JP7159270B2 - Methods and procedures for non-invasive evaluation of genetic mutations - Google Patents

Description

RELATED PATENT APPLICATIONS This patent application was filed on October 4, 2013, is entitled "METHODS AND PROCESSES FOR NON-INVASIVE ASSESSMENT OF GENETIC VARIATIONS", lists Gregory Hannum as the inventor, and has SEQ as the docket number. -6073-PV, which claims the benefit of US Provisional Patent Application No. 61/887,081. The entire contents of the above application, including all text, tables and drawings, are incorporated herein by reference.

The technology provided herein relates, in part, to methods, processes and machines for the non-invasive assessment of genetic mutations.

Genetic information in living organisms (eg, animals, plants and microorganisms) and other forms of replicating genetic information (eg, viruses) is encoded in deoxyribonucleic acid (DNA) or ribonucleic acid (RNA). Genetic information is contiguous or modified nucleotides, which describe the chemical or hypothetical primary structure of a nucleic acid. For humans, the complete genome contains approximately 30,000 genes located on 24 chromosomes (see Non-Patent Document 1). Each gene encodes a specific protein that performs a specific biochemical function after expression through transcription and translation in a living cell.

Many medical conditions are caused by mutations in one or more genes. Mutations in certain genes cause medical conditions such as hemophilia, thalassemia, Duchenne muscular dystrophy (DMD), Huntington's disease (HD), Alzheimer's disease and cystic fibrosis (CF). (Non-Patent Document 2). Such genetic diseases can result from single nucleotide additions, substitutions or deletions in the DNA of particular genes. For example, certain congenital defects are associated with chromosomal abnormalities, also called aneuploidies, such as trisomy 21 (Down's syndrome), trisomy 13 (Patou's syndrome), trisomy 18 (Edwards' syndrome), trisomy 16 and trisomy 22, monosomy X. (Turner's syndrome), and certain sex chromosome aneuploidies, such as Klinefelter's syndrome (XXY). Another genetic variation is fetal sex, which can often be determined based on the X and Y of the sex chromosomes. Mutations in several genes lead to several diseases such as diabetes, arteriosclerosis, obesity, various autoimmune diseases and cancers (e.g. colorectal cancer, breast cancer, ovarian cancer, lung cancer). Either way, the individual may become susceptible or may develop such a disease.

The Human Genome, T. Strachan, BIOS Scientific Publishers, 1992 Human Genome Mutations, D. N. Cooper and M. Krawczak, BIOS Publishers, 1993

Identification of mutations or variances in one or more genes can lead to diagnosis of particular medical conditions or determination of predisposition to such conditions. Identification of genetic variance can facilitate medical decision-making and/or access useful medical procedures. In certain embodiments, identifying mutations or variances in one or more genes comprises analysis of cell-free DNA. Cell-free DNA (CF-DNA) is composed of DNA fragments that arise from cell death and circulate in peripheral blood. High concentrations of CF-DNA can be indicative of certain clinical conditions such as cancer, trauma, burns, myocardial infarction, stroke, sepsis, infection and other diseases. Additionally, cell-free fetal DNA (CFF-DNA) can be detected in the maternal bloodstream and used for various non-invasive prenatal diagnostic methods.

In certain aspects herein, a system comprising a memory and one or more microprocessors, wherein the one or more microprocessors read biases for sequences of a sample according to instructions in the memory. (a) of (i) a local genomic bias estimate and (ii) a bias frequency for the sequence reads of the test sample generating a relationship, thereby generating a sample-biased relationship, wherein the sequence reads are of circulating cell-free nucleic acids derived from the test sample and mapped against a reference genome; b) comparing the sample bias relationship and the reference bias relationship, thereby generating a comparison, comprising:
The reference bias relationship is between (i) the local genomic bias estimate and (ii) the bias frequency for the reference, and (c) the sample's normalizing the sequence read counts, wherein the sequence read bias with respect to the sample is reduced.

In certain aspects herein, a system comprising a memory and one or more microprocessors, wherein the one or more microprocessors are configured to bias read sequences relative to a sample according to instructions in the memory. (a) the relationship between (i) guanine and cytosine (GC) density and (ii) GC density frequency for the sequence reads of the test sample generating, thereby generating a sample GC density relationship, wherein the sequence reads are of circulating cell-free nucleic acids derived from the test sample, and the sequence reads are mapped against a reference genome and (b) comparing the sample GC density relation and the reference GC density relation, thereby generating a comparison, wherein the reference GC density relation is the (i) GC density for the reference and (ii and (c) normalizing the sequence read counts for the sample according to the comparison determined in (b), thereby reducing sequence read bias for the sample. A system is provided, comprising the steps of:

In certain aspects herein, a system comprising a memory and one or more microprocessors, wherein the one or more microprocessors detect the presence of aneuploidy for a sample according to instructions in the memory. or non-existent, wherein the processing comprises: (a) filtering a portion of the reference genome according to the read density distribution, whereby a test sample comprising read densities of the filtered portion; wherein the read density comprises reads of sequences of circulating cell-free nucleic acids derived from test samples derived from pregnant females, the read density distribution comprising for the plurality of samples (b) adjusting the read density profile for the test sample according to one or more principal components obtained from a set of known euploid samples by principal component analysis; (c) comparing the test sample profile with the reference profile, thereby providing a comparison; (d) chromosomes for the test sample according to the comparison; and determining the presence or absence of aneuploidy.

Certain aspects of this technology are further described in the following description, examples, claims and drawings.

The drawings illustrate and do not limit embodiments of the technology. For clarity and clarity of illustration, the drawings are not drawn to scale, and in some instances various aspects may be exaggerated or enlarged to facilitate understanding of certain embodiments. may be indicated as

FIG. 1 shows an embodiment of the GC density obtained with the Epanechnikov kernel (bandwidth=200 bp).

FIG. 2 shows a plot of GC density (y-axis) for the HTRA1 gene, where GC density is normalized across the genome. Genomic location is indicated on the x-axis.

FIG. 3 shows the distribution of local genomic bias (eg, GC density, x-axis) estimates for the reference genome (solid line) and sequence reads from the sample (dashed line). Biased frequencies (eg, density frequencies) are shown on the y-axis. Estimates of GC density are normalized across the genome. In this example, the sample has more reads with higher GC content than expected from the reference.

FIG. 4 shows a comparison of the distribution of the GC density estimates for the reference genome and the sequence read GC density estimates for the sample using a weighted cubic polynomial fit relationship. Estimates of GC density (x-axis) were normalized across the genome. The GC density frequency is expressed on the y-axis as the logarithm (log ₂ ) of the ratio of the density frequency of the reference divided by the density frequency of the sample.

FIG. 5A shows the distribution of median GC density (x-axis) for all parts of the genome. FIG. 5B shows the median absolute deviation (MAD) (x-axis) determined by the GC density distribution for multiple samples. GC density frequency is shown on the y-axis. Each portion was filtered by the distribution of the median GC density for multiple reference samples (eg, training set), and the MAD value was determined by the GC density distribution of multiple samples. Portions containing GC densities that fell outside an established threshold (eg, four times the interquartile range of the MAD) were filtered out from consideration.

FIG. 6A shows the sample read density profile for the genome, including the median read density (y-axis, eg, read density/portion), and the relative position of each genome portion within the genome (x-axis, portion index). FIG. 6B shows the first principal component (PC1) and FIG. 6C shows the second principal component (PC2) obtained from principal component analysis of read density profiles obtained from a training set of 500 euploids. indicates

Figures 7A-C show examples of sample read density profiles for genomes containing trisomy of chromosome 21 (eg, between two vertical lines). The relative position of each genome part is indicated on the x-axis. Read density is presented on the y-axis. FIG. 7A shows a raw (eg, unadjusted) read density profile. FIG. 7B shows the profile of 7A with a first adjustment involving subtraction of the median profile. FIG. 7C shows the profile of 7B with the second adjustment. The second adjustment involves subtraction of the 8-fold principal component profile, weighted based on representations found in this sample (eg, model built). For example, SampleProfile=A*PC1+B*PC2+C*PC3 .

FIG. 8 shows QQ-plots of test p-values from bootstrapped training samples for the T21 test. A QQ plot generally compares two distributions. FIG. 8 shows a comparison of the ChAI scores (y-axis) from the test samples with a uniform distribution (ie, distribution of expected p-values, x-axis). Each point represents the log-p value score of a single test sample. The samples are sorted and assigned an "expected" value (x-axis) based on uniform distribution. The lower dashed line represents the diagonal line and the upper line represents the Bonferroni threshold. Samples with a uniform distribution are expected to lie on the lower diagonal (lower dashed line). Data values are far off the diagonal (eg, bias) due to intra-part correlations, indicating samples with higher scoring (lower p-values) than expected. The methods described herein (eg ChAI, see eg Example 1) can correct for this observed bias.

FIG. 9A shows read density plots showing differences in PC2 coefficients for males and females in the training set. FIG. 9B shows a Receiver Operating Characteristic (ROC) plot for gender determination using the PC2 coefficient. Gender determinations made by sequencing were used as authenticity references.

10A-10B illustrate system embodiments. 10A-10B illustrate system embodiments.

FIG. 11 shows one embodiment of the system.

FIG. 12 illustrates one embodiment of the methods provided herein.

Next-generation sequencing makes it possible to sequence nucleic acids on a genome-wide scale by methods that are faster and cheaper than traditional methods of sequencing. The methods, systems, and products provided herein can leverage advanced sequencing technology to locate and identify genetic mutations and/or associated diseases and disorders. The methods, systems, and products provided herein can often provide non-invasive assessment of a subject genome (e.g., a fetal genome) using blood samples or portions thereof, and often more invasive methods. safer, faster, and/or cheaper than sexual techniques (eg, amniocentesis, biopsy). In some embodiments, as used herein, in part, obtaining sequence reads of nucleic acids present in a sample, wherein the sequence reads are often mapped against a reference sequence; , a method comprising the steps of processing sequence read counts and determining the presence or absence of genetic mutations. The systems, methods, and articles of manufacture provided herein are useful for locating and/or identifying genetic mutations, and diagnosing diseases, disorders, and disabilities associated with certain genetic mutations. and useful for treating

Also provided herein, in some embodiments, are data manipulation methods for reducing and/or removing sequencing biases introduced by various aspects of sequencing technology. Sequencing bias often contributes to uneven distribution of reads across the genome or segments thereof, and/or variation in read quality. Sequencing bias can corrupt genomic sequencing data, impair valid data analysis, contaminate results, and prevent accurate data interpretation. Sometimes sequencing bias can be reduced by increasing sequencing coverage, but this approach often inflates sequencing costs and has very limited effectiveness. The data manipulation methods described herein can reduce and/or eliminate sequencing bias, thereby improving the quality of sequence read data without increasing sequencing costs. Also provided herein, in some embodiments, are systems, machines, apparatus, articles of manufacture, and modules that implement the methods described herein.

Samples Provided herein are methods and compositions for analyzing nucleic acids. In some embodiments, nucleic acid fragments in a mixture of nucleic acid fragments are analyzed. A mixture of nucleic acids can have two or more different nucleotide sequences, different fragment lengths, different origins (e.g., genomic origin, fetal vs. maternal origin, cellular or tissue origin, sample origin, subject origin, etc.), or combinations thereof. More nucleic acid fragment species can be included.

Nucleic acids or nucleic acid mixtures utilized in the methods, systems, machines, and/or devices described herein are often isolated from a sample obtained from a subject (eg, a test subject). A subject from which a specimen or sample is obtained is sometimes referred to herein as a test subject. A subject can be any living or non-living organism, including, but not limited to, humans, non-human animals, plants, bacteria, fungi, viruses or protists. Mammals, reptiles, birds, amphibians, fish, ungulates, ruminants, bovines (e.g. cattle), equines (e.g. horses), caprines and ovines ( sheep, goats), swine (e.g. pigs), camelids (e.g. camels, llamas, alpacas), monkeys, apes (e.g. gorillas, chimpanzees), bears (e.g. bears), poultry, Any human or non-human animal can be selected, including dogs, cats, mice, rats, fish, dolphins, whales and sharks. A subject can be male or female (eg, female, pregnant, pregnant female). A subject can be of any age (eg, embryo, fetus, infant, child, adult).

Nucleic acids can be isolated from any type of suitable biological specimen or sample (eg, test sample). A sample or test sample can be any specimen isolated or obtained from a subject or portion thereof (eg, a human subject, pregnant female, fetus). A test sample is often obtained from a test subject. Test samples are often obtained from pregnant females (eg, pregnant human females). Non-limiting examples of analytes include bodily fluids or tissues obtained from a subject, including, but not limited to, blood or blood products (eg, serum, plasma, etc.), cord blood, chorionic villi , amniotic fluid, cerebrospinal fluid, spinal fluid, lavage fluids (e.g., from bronchoalveolar, stomach, peritoneal, ductal, ear, arthroscopy), biopsies (e.g., from preimplantation embryos). samples), peritoneal puncture samples, cells (blood cells, placental cells, embryonic or fetal cells, embryonic nucleated cells or embryonic cell remnants) or parts thereof (e.g., mitochondria, nuclei, extracts, etc.) , female reproductive system washes, urine, feces, sputum, saliva, nasal mucus, prostatic fluid, lavage, semen, lymph, bile, tears, sweat, milk, mammary fluid, etc., or combinations thereof. A test sample may comprise blood or blood products (eg, plasma, serum, lymphocytes, platelets, buffy coat). Test samples sometimes include serum obtained from pregnant females. Test samples sometimes include plasma obtained from pregnant females. In some embodiments, the biological sample is a cervical swab obtained from the subject. In some embodiments, the biological sample can be blood, and sometimes plasma or serum. The term "blood," as used herein, refers to a blood sample or preparation from a subject (eg, a test subject, eg, a pregnant woman or a woman being tested for possible pregnancy). The term includes whole blood, blood products or any fraction of blood, including serum, plasma, buffy coat, etc. according to conventional definitions. Blood or a fraction thereof often contains nucleosomes (eg, maternal and/or fetal nucleosomes). Nucleosomes contain nucleic acids and are sometimes cell-free or intracellular nucleosomes. Blood also includes a buffy coat. The buffy coat is sometimes isolated by utilizing a Ficoll gradient. A buffy coat can contain white blood cells (eg, white blood cells, T cells, B cells, platelets, etc.). In certain embodiments, the buffy coat comprises maternal and/or fetal nucleic acids. Plasma refers to the fraction of whole blood obtained as a result of centrifugation of anticoagulant-treated blood. Serum refers to the aqueous liquid portion that remains after a blood sample has clotted. Fluid or tissue samples are often collected according to standard protocols commonly followed by hospitals or outpatient clinics. In the case of blood, an appropriate volume of peripheral blood (eg, 3-40 milliliters) is often collected and can be stored according to standard procedures before or after preparation. A bodily fluid or tissue sample from which nucleic acids are extracted may be cell-free (eg, cell-free). In some embodiments, a bodily fluid or tissue sample may contain cellular elements or cellular remnants. In some embodiments, fetal or cancerous cells may be included in the sample.

Often the sample is heterogeneous, meaning that more than one type of nucleic acid species is present in the sample. For example, heterogeneous nucleic acids include, but are not limited to, (i) fetal and maternally derived nucleic acids, (ii) cancerous and non-cancerous nucleic acids, (iii) pathogenic and host nucleic acids. and more generally (iv) mutated and wild-type nucleic acids. A sample may be heterogeneous, as more than one cell type is present, such as fetal and maternal cells, cancerous and non-cancerous cells, or pathogen and host cells. is. In some embodiments, minor and major nucleic acid species are present.

When applying the techniques described herein prenatally, body fluids or tissue samples can be collected from females at the appropriate gestational age to be tested or from females being tested for fertility potential. can. The appropriate gestational age may vary depending on the prenatal examination being performed. In certain embodiments, the pregnant female subject is sometimes in the first trimester, sometimes in the second trimester, or sometimes in the third trimester. In certain embodiments, bodily fluids or tissues are obtained from pregnant females from about 1 to about 45 weeks of gestation (eg, 1-4, 4-8, 8-12, 12-16, 16-20 weeks of gestation). , 20-24, 24-28, 28-32, 32-36, 36-40 or 40-44 weeks), sometimes from about 5 to about 28 weeks gestation (e.g., 6, 7, 8, 9 , 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26 or 27 weeks). In certain embodiments, a body fluid or tissue sample is obtained from a pregnant female during or shortly after (eg, 0-72 hours) birth (eg, vaginal or non-vaginal delivery (eg, surgical delivery)). later).

Obtaining Blood Samples and Extracting DNA The methods herein are often used during and sometimes after pregnancy to detect the presence or absence of mutations in maternal and/or fetal genes and/or fetal and/or gestational cancers. Non-invasive means for monitoring the health of females in utero include the isolation, enrichment and analysis of fetal DNA found in maternal blood. Accordingly, the first steps in carrying out certain methods herein often involve obtaining a blood sample from a pregnant woman and extracting DNA from the sample.

Obtaining Blood Samples Blood samples can be obtained from pregnant women at the appropriate gestational age to be tested using methods according to the present technology. Appropriate gestational age can vary depending on the disorder being tested, as discussed below. Collection of blood from women is often performed according to standard protocols commonly followed by hospitals or outpatient clinics. Appropriate amounts of peripheral blood, typically 5-50 ml, can often be collected and stored according to standard procedures before further preparation. A blood sample can be collected, stored, or shipped in a manner that minimizes degradation of the quality of nucleic acids present in the sample.

Preparation of Blood Samples Analysis of fetal DNA found in maternal blood can be performed using, for example, whole blood, serum or plasma. Methods for preparing serum or plasma from maternal blood are known. For example, a pregnant woman's blood can be placed in tubes containing EDTA or a special commercial product, such as Vacutainer SST (Becton Dickinson, Franklin Lakes, N.J.) to prevent blood clotting, and then Plasma can be obtained from whole blood by centrifugation. Serum can be obtained with or without centrifugation after blood clotting. When using centrifugation, it is typically, but not necessarily, carried out at a suitable speed, eg, 1,500-3,000 g. Plasma or serum may undergo an additional centrifugation step before being transferred to a new tube for DNA extraction.

In addition to the cell-free portion of whole blood, DNA can also be recovered from the cell fraction and enriched in the buffy coat portion, which is a whole blood sample obtained from a woman. It can be obtained by centrifuging and removing plasma.

DNA Extraction There are a number of known methods for extracting DNA from biological samples, including blood. General methods of DNA preparation (eg, as described by Sambrook and Russell, Molecular Cloning: A Laboratory Manual 3d ed. 2001) can be followed and various commercially available reagents or kits, such as those from Qiagen. QIAamp Circulating Nucleic Acid Kit, QiaAmp DNA Mini Kit, or QiaAmp DNA Blood Mini Kit (Qiagen, Hilden, Germany); DNA can also be obtained from blood samples obtained from pregnant women using the Purification Kit (Amersham, Piscataway, N.J.). Combinations of more than one of these methods can also be used.

In some embodiments, the sample may first be enriched or to some extent enriched for fetal nucleic acids by one or more methods. For example, the compositions and processes of the present technology can be used alone or in combination with other discriminating agents to discriminate between fetal and maternal DNA. Examples of these factors include single-nucleotide differences between the X and Y chromosomes, sequences specific to the Y chromosome, polymorphisms located elsewhere in the genome, and fetal and maternal DNA. and differences in methylation patterns between maternal and fetal tissue.

Other methods for enriching a sample for particular species of nucleic acids are described in PCT Patent Application No. PCT/US07/69991 filed May 30, 2007, PCT Patent Application No. PCT/ US 2007/071232, US Provisional Application Nos. 60/968,876 and 60/968,878 (assigned to Applicant) (PCT Patent Application No. PCT/EP05/012707 filed Nov. 28, 2005) , all of which are incorporated herein by reference. In certain embodiments, maternal nucleic acids are selectively (partially, substantially, almost completely, or completely) removed from the sample.

The terms "nucleic acid" and "nucleic acid molecule" can be used interchangeably throughout this disclosure. These terms include DNA (e.g., complementary DNA (cDNA), genomic DNA (gDNA), etc.), RNA (e.g., messenger RNA (mRNA), small interfering RNA (siRNA), ribosomal RNA (rRNA), tRNA, microRNA, fetal or placental highly expressed RNA, etc.), and/or analogs of DNA or RNA (such as those containing base analogs, sugar analogs and/or exogenously added backbones, etc.), RNA /DNA hybrids and polyamide nucleic acids (PNA), etc., all of which may be in single- or double-stranded form and, unless otherwise limited, are naturally occurring Known analogues of natural nucleotides can be included that can function in a manner analogous to the nucleotides in the base. In certain embodiments, the nucleic acid is a plasmid, phage, autonomously replicating sequence (ARS), centromere, artificial chromosome, chromosome, or replicating in vitro or in a host cell, cell, cell nucleus, or cytoplasm of a cell. It can also be other nucleic acids that can be obtained or replicated, or can be derived from them. A template nucleic acid, in some embodiments, can be derived from a single chromosome (eg, a nucleic acid sample can be derived from one chromosome of a sample obtained from a diploid organism). Unless otherwise limited, the term encompasses nucleic acids containing known analogues of natural nucleotides that have similar binding properties to the reference nucleic acid and are metabolized in a manner similar to naturally occurring nucleotides. Unless otherwise specified, a particular nucleic acid sequence includes not only the explicitly indicated sequence, but also conservatively modified variants thereof (e.g., degenerate codon substitutions), alleles, orthologs, single nucleotide polymorphisms (SNPs) and complementary sequences are also implicitly included. Specifically, by generating a sequence in which the third position of one or more selected (or all) codons is replaced with a mixed base residue and/or a deoxyinosine residue. , degenerate codon substitutions can be obtained. The term nucleic acid is used interchangeably with locus, gene, cDNA, and mRNA that the gene encodes. The term also includes, as equivalents, derivatives, variants and analogs of RNA or DNA synthesized from analogs of nucleotides, single strands (“sense” or “antisense” strands, “plus” or “minus” strands). , "forward" or "reverse" reading frames), and double-stranded polynucleotides. The term "gene" refers to a segment of DNA involved in the production of a polypeptide chain, including regions preceding and following the coding region involved in transcription/translation and regulation of transcription/translation of the gene product. (leader and trailer), as well as intervening sequences (introns) between individual coding segments (exons).

Deoxyribonucleotides include deoxyadenosine, deoxycytidine, deoxyguanosine, and deoxythymidine. For RNA, the base cytosine is replaced with uracil. A template nucleic acid can be prepared using a nucleic acid obtained from a subject as a template.

Nucleic Acid Isolation and Processing Nucleic acids can be obtained from one or more sources (eg, cells, serum, plasma, buffy coat, lymph, skin, soil, etc.) by methods known in the art. Nucleic acids are often isolated from test samples. Any suitable method can be used to isolate, extract and/or purify DNA from a biological sample (e.g. blood or blood products), non-limiting examples of which are: Methods of DNA preparation (e.g., described by Sambrook and Russell, Molecular Cloning: A Laboratory Manual 3d ed. 2001), various commercially available reagents or kits, e.g., Qiagen's QIAamp Circulating Nucleic Acid Kit, QiaAmp DNA Mini Kit, or QiaAmp DNA Blood Mini Kit (Qiagen, Hilden, Germany), GenomicPrep™ Blood DNA Isolation Kit (Promega, Madison, Wis.), and GFX™ Genomic Blood DNA Purification Kit (America, Persia, Persia). .J.), etc., or combinations thereof.

Cell lysis procedures and reagents are known in the art and generally include chemical methods (e.g., detergents, hypotonic solutions, enzymatic procedures, etc., or combinations thereof), physical methods (e.g., French press, ultrasonic treatment, etc.), or a dissolution method using an electrolyte. Any suitable lysis procedure can be utilized. For example, chemical methods generally utilize lysing agents to disrupt cells and extract nucleic acids from cells, followed by treatment with chaotropic salts. Physical methods such as freeze/thaw followed by comminution; use of cell presses and the like are also useful. Lysis procedures with high salt concentrations are also commonly used. For example, an alkaline lysis procedure can be used. The latter procedure traditionally incorporates the use of a phenol-chloroform solution, and an alternative phenol-chloroform-free procedure involving three solutions is also available. For the latter procedure, one solution can contain 15 mM Tris, pH 8.0; 10 mM EDTA, and 100 μg/ml ribonuclease A; a second solution contains 0.2 N NaOH and 1% SDS. a third solution can contain 3M KOAc, pH 5.5. These procedures are described in Current Protocols in Molecular Biology, John Wiley & Sons, N.W. Y. , 6.3.1-6.3.6 (1989), which is incorporated herein in its entirety.

The nucleic acid, when compared to another nucleic acid, can be isolated at different time points and each sample is derived from the same or different sources. For example, a nucleic acid can be derived from a nucleic acid library, such as a cDNA library or an RNA library. Nucleic acids may be the result of purification or isolation of nucleic acids and/or amplification of nucleic acid molecules obtained from a sample. Nucleic acid provided for the processes described herein may be nucleic acid from one sample, or from two or more samples (e.g., one or more, two or more, three or more, 4 or more, 5 or more, 6 or more, 7 or more, 8 or more, 9 or more, 10 or more, 11 or more, 12 or more, 13 or more, 14 or more, 15 or more, 16 or more, 17 or more, 18 or more, 19 or more, or 20 or more sample).

In certain embodiments, nucleic acids can include extracellular nucleic acids. The term "extracellular nucleic acid," as used herein, can refer to nucleic acid isolated from a source that is substantially free of cells, and can also refer to "cell-free" nucleic acid and/or "cell-free" nucleic acid. Also referred to as "cell cycling" nucleic acids. Extracellular nucleic acids are present in and can be obtained from blood (eg, the blood of pregnant females). Extracellular nucleic acids are often free of detectable cells and may contain cellular elements or cellular remnants. Non-limiting examples of cell-free sources for obtaining extracellular nucleic acids are blood, plasma, serum and urine. As used herein, the term "obtaining a cell-free circulating sample nucleic acid" refers to obtaining a sample directly (e.g., collecting a sample, e.g., a test sample) or including obtaining Without being limited by theory, extracellular nucleic acids may be products of cell apoptosis and cell degradation, which give rise to extracellular nucleic acids that often have a series of lengths spanning a spectrum (e.g., a "ladder"). .

In certain embodiments, extracellular nucleic acid can comprise different nucleic acid species, and is therefore referred to herein as "heterogeneous." For example, serum or plasma obtained from a person with cancer may contain nucleic acid derived from cancerous cells and nucleic acid derived from non-cancerous cells. In another example, serum or plasma obtained from a pregnant female may contain maternal and fetal nucleic acids. In some cases, the fetal nucleic acid is sometimes about 5% to about 50% of the total nucleic acid (eg, about 4, 5, 6, 7, 8, 9, 10, 11, 12, 13 of the total nucleic acid). , 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38 , 39, 40, 41, 42, 43, 44, 45, 46, 47, 48 or 49% is fetal nucleic acid). In some embodiments, the length of the majority of the fetal nucleic acid in the nucleic acid is about 500 base pairs or less, about 250 base pairs or less, about 200 base pairs or less, about 150 base pairs or less , about 100 base pairs or less, about 50 base pairs or less, or about 25 base pairs or less.

In certain embodiments, the nucleic acid can be provided and the methods described herein performed without processing the nucleic acid-containing sample. In some embodiments, a sample containing nucleic acids is processed prior to providing the nucleic acids and performing the methods described herein. For example, nucleic acids can be extracted, isolated, purified, partially purified, or amplified from a sample. The term "isolated," as used herein, refers to removing a nucleic acid from its original environment (e.g., the natural environment when naturally occurring or the host cell when exogenously expressed). Thus, a nucleic acid has been altered in that it has been removed from its original environment by human intervention (eg, "by the hand of man"). The term "isolated nucleic acid," as used herein, can refer to nucleic acid that has been removed from a subject (eg, a human subject). An isolated nucleic acid may be provided with less non-nucleic acid components (eg, proteins, lipids) than the amount of components present in the source sample. A composition comprising an isolated nucleic acid may be from about 50% to greater than 99% free of non-nucleic acid components. A composition comprising an isolated nucleic acid is free of more than about 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99% or 99% of it free of non-nucleic acid components Sometimes. The term "purified," as used herein, provides a nucleic acid that contains less non-nucleic acid components (e.g., proteins, lipids, carbohydrates) than the amount of non-nucleic acid components (e.g., proteins, lipids, carbohydrates) present prior to subjecting the nucleic acid to a purification procedure. can point to Compositions comprising purified nucleic acid are about 80%, 81%, 82%, 83%, 84%, 85%, 86%, 87%, 88%, 89%, 90%, 91%, 92%, 93% thereof. , 94%, 95%, 96%, 97%, 98%, 99% or more than 99% may not contain other non-nucleic acid components. The term "purification," as used herein, can refer to providing a nucleic acid that contains fewer nucleic acid species than in the sample source from which the nucleic acid was derived. Compositions comprising purified nucleic acids are free of other nucleic acid species in greater than about 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99% or 99% Sometimes. For example, fetal nucleic acid can be purified from a mixture comprising maternal and fetal nucleic acid. In certain instances, nucleosomes containing small fragments of fetal nucleic acid can be purified from a mixture of larger nucleosome complexes containing larger fragments of maternal nucleic acid.

In some embodiments, nucleic acids are fragmented or cleaved before, during, or after the methods described herein. Fragmented or truncated nucleic acids can be from about 5 to about 10,000 base pairs, from about 100 to about 1,000 base pairs, from about 100 to about 500 base pairs, or from about 10, 15, 20, 25, 30, 35, 40, 45, 50, 55, 60, 65, 70, 75, 80, 85, 90, 95, 100, 200, 300, 400, 500, 600, 700, 800, 900, 1000, 2000, 3000, 4000, It can have a nominal average or mean length of 5000, 6000, 7000, 8000 or 9000 base pairs. Fragments can be produced by any suitable method known in the art, and the average, average or nominal length of nucleic acid fragments can be controlled by choosing an appropriate fragment production procedure.

Nucleic acid fragments can contain overlapping nucleotide sequences, and such overlapping sequences can facilitate construction of the unfragmented nucleotide sequence of the corresponding nucleic acid, or a segment thereof. For example, one fragment may have subsequences x and y and another fragment may have subsequences y and z, where x, y and z may be 5 or more nucleotides in length. Nucleotide sequence. In certain embodiments, overlapping sequences y can be utilized to facilitate assembly of xyz nucleotide sequences in nucleic acids derived from a sample. In certain embodiments, nucleic acids may be partially fragmented (eg, from an incomplete or truncated specific cleavage reaction) or fully fragmented.

In some embodiments, nucleic acids are fragmented or cleaved by a suitable method, non-limiting examples of which include physical methods (e.g., shearing, e.g., sonication, French press, heating, UV irradiation). etc.), enzymatic treatments (e.g. enzymatic cleaving agents (e.g. suitable nucleases, suitable restriction enzymes, suitable methylation-sensitive restriction enzymes)), chemical methods (e.g. alkylation, DMS, piperidine, acid hydrolysis, base hydrolysis, heating, etc., or combinations thereof), treatments described in US Patent Application Publication No. 20050112590, etc., or combinations thereof.

As used herein, "fragmentation" or "cleavage" can split a nucleic acid molecule, such as a nucleic acid template gene molecule or its amplification product, into two or more smaller nucleic acid molecules. Refers to a procedure or condition. Such fragmentation or cleavage can be sequence-specific, base-specific, or non-specific, among a variety of methods, reagents or conditions, including, for example, chemical, enzymatic, physical fragmentation. can be achieved by either

As used herein, "fragment," "cleavage product," "cleaved product," or grammatical variations thereof, refer to nucleic acid molecules obtained as a result of fragmentation or cleavage of a nucleic acid template gene molecule. , or their amplification products. Although such fragments or cleaved products may refer to all nucleic acid molecules resulting from a cleavage reaction, typically such fragments or cleaved products are nucleic acid template gene molecules. It refers only to nucleic acid molecules obtained as a result of fragmentation or cleavage of a nucleic acid template gene molecule or amplification product segments thereof that contain the corresponding nucleotide sequence of . The term "amplification," as used herein, refers to the linear or exponential generation of a target nucleic acid in a sample into amplicon nucleic acids having the same or substantially the same nucleotide sequence as the target nucleic acid or segment thereof. Refers to processing. In certain embodiments, the term "amplification" refers to methods involving the polymerase chain reaction (PCR). For example, an amplification product can contain one or more more nucleotides than the amplified nucleotide region of the nucleic acid template sequence (e.g., the primers, in addition to the nucleotides complementary to the nucleic acid template gene molecule, have "extra can contain "extra" nucleotides, such as the transcription initiation sequence, resulting in an amplification product containing "extra" nucleotides or nucleotides that do not correspond to the amplified nucleotide region of the nucleic acid template gene molecule. ). Thus, a fragment can include a fragment resulting from a segment or portion of an amplified nucleic acid molecule that contains, at least in part, nucleotide sequence information obtained from or based on a displayed nucleic acid template molecule.

As used herein, the term "complementary cleavage reactions" refers to cleavage reactions performed on the same nucleic acid using different cleavage reagents or by varying the cleavage specificity of the same cleavage reagent. , thus generating alternative cleavage patterns of the same target or reference nucleic acid or protein. In certain embodiments, nucleic acids are treated with one or more specific cleaving agents (e.g., 1, 2, 3, 4, 5, 6, 7, 8, 9, 10 or more specific cleaving agents) (eg, nucleic acids are treated with each specific cleaving agent in separate vessels). The term "specific cleaving agent" as used herein refers to an agent, sometimes a chemical or an enzyme, capable of cleaving a nucleic acid at one or more specific sites.

Also, the nucleic acid can be exposed to treatments that modify certain nucleotides in the nucleic acid prior to providing the nucleic acid to the methods described herein. For example, a process can be applied to a nucleic acid that selectively modifies the nucleic acid based on the methylation status of nucleotides therein. In addition, conditions such as high temperature, ultraviolet radiation, X-radiation can induce changes in the sequence of nucleic acid molecules. Nucleic acids can be provided in any suitable form useful for conducting suitable sequence analyses.

Nucleic acids may be single-stranded or double-stranded. For example, single-stranded DNA can be generated by denaturing double-stranded DNA, eg, by treatment with heat or alkali. In certain embodiments, the nucleic acid adopts a D-loop structure formed by inserting an oligonucleotide into the strand of a double-stranded DNA molecule, or in a DNA-like molecule, such as a peptide nucleic acid (PNA). be. Formation of the D-loop is described by E. It can be aided by adding E. coli RecA protein and/or changing the salt concentration, eg, using methods known in the art.

Determining Fetal Nucleic Acid Content In some embodiments, the amount (eg, concentration, relative amount, absolute amount, copy number, etc.) of fetal nucleic acid in a nucleic acid is determined. In certain embodiments, the amount of fetal nucleic acid in a sample is referred to as the "fetal fraction." In some embodiments, "fetal fraction" refers to the fraction of fetal nucleic acid in circulating cell-free nucleic acid in a sample (eg, blood sample, serum sample, plasma sample) obtained from a pregnant female. In certain embodiments, markers specific for male fetuses (e.g., Y-chromosome STR markers (e.g., DYS19, DYS385, DYS392 markers); RhD markers in RhD-negative females), alleles of polymorphic sequences, according to the ratio, or one or more markers that are specific for fetal nucleic acid and not maternal nucleic acid (e.g., epigenetic biomarker differences between mother and fetus (e.g., methylation; see further below). described in detail), or fetal RNA markers in maternal plasma (see, e.g., Lo, 2005, Journal of Histochemistry and Cytochemistry, 53(3):293-296)). Determine the amount of nucleic acid.

Determination of fetal nucleic acid content (e.g., fetal fraction) is sometimes performed by a fetal quantifier assay, e.g., as described in US Patent Application Publication No. 2010/0105049, which is incorporated herein by reference. (FQA). This type of assay allows fetal nucleic acid in a maternal sample to be detected and quantified based on the methylation status of nucleic acids in the sample. In certain embodiments, the amount of fetal nucleic acid derived from a maternal sample can be determined relative to the total amount of nucleic acid present, thereby yielding the percentage of fetal nucleic acid in the sample. In certain embodiments, the copy number of fetal nucleic acid in a maternal sample can be determined. In certain embodiments, in a sequence-specific (or partly-specific) manner, sometimes to allow accurate chromosome dose analysis (e.g., to detect the presence or absence of fetal aneuploidy). With sufficient sensitivity, the amount of fetal nucleic acid can be determined.

A fetal quantification assay (FQA) can be performed in conjunction with any of the methods described herein. Any art-known method and/or as described in U.S. Patent Application Publication No. 2010/0105049, for example, to distinguish between maternal and fetal DNA based on differences in methylation status, Such assays can be performed by any method that can quantify (eg, determine the amount of) DNA. Methods for differentiating nucleic acids based on methylation status include, but are not limited to, by methylation sensitivity, e.g. (MBD-FC)) (Gebhard et al. (2006) Cancer Res. 66(12):6118-28); methylation-specific antibodies; MSP (methylation-sensitive PCR), COBRA, extension of primers by methylation-sensitive single nucleotides (Ms-SNuPE), or Sequenom MassCLEAVE™ technology; and use of methylation-sensitive restriction enzymes (e.g., maternal are digested using one or more methylation-sensitive restriction enzymes, thereby enriching for fetal DNA). Methyl-sensitive enzymes can also be used to differentiate nucleic acids on the basis of their methylation status, and these enzymes will, for example, favor their DNA recognition sequences if the latter are unmethylated. Cleavage or digestion can be performed either systematically or substantially. Therefore, unmethylated DNA samples are cut into smaller pieces than methylated DNA samples, and hypermethylated DNA samples are not cut. Unless explicitly stated, any method for differentiating nucleic acids based on methylation status can be used with the compositions and methods of the techniques herein. The amount of fetal DNA can be determined, for example, by introducing one or more competitors at known concentrations during the amplification reaction. Determination of the amount of fetal DNA can also be performed, for example, by RT-PCR, primer extension, sequencing and/or counting. In certain instances, the amount of nucleic acid can be determined using BEAMing technology as described in US Patent Application Publication No. 2007/0065823. In certain embodiments, the restriction efficiency can be determined and the efficiency ratio is used to further determine the amount of fetal DNA.

In certain embodiments, a fetal quantification assay (FQA) can be used to determine the concentration of fetal DNA in a maternal sample by, for example: a) present in the maternal sample determining the total amount of DNA; b) selectively digesting maternal DNA in a maternal sample using one or more methylation-sensitive restriction enzymes, thereby enriching fetal DNA; c) determining the amount of fetal DNA obtained from step b); d) comparing the amount of fetal DNA obtained from step c) with the total amount of DNA obtained from step a), whereby maternal Determine the concentration of fetal DNA in the sample. In certain embodiments, the absolute copy number of fetal nucleic acids in a maternal sample can be determined using, for example, mass spectrometry and/or a system that uses competitive PCR approaches to measure absolute copy number. can. See, for example, Ding and Cantor (2003) PNAS, USA 100:3059-3064, and US Patent Application Publication No. 2004/0081993, both of which are incorporated herein by reference.

In certain embodiments, based on allelic ratios of polymorphic sequences (e.g., single nucleotide polymorphisms (SNPs)), for example, US Patent Application Publication No. 2011/0224087, incorporated herein by reference. The fetal fraction can be determined using methods such as those described in pp. In such methods, nucleotide sequence reads are obtained for a maternal sample, and nucleotide sequence reads that map to a first allele at an informative polymorphic site (e.g., a SNP) in a reference genome. with the total number of nucleotide sequence reads that map to the second allele to determine the fetal fraction. In certain embodiments, e.g., for a mixture of fetal and maternal nucleic acids in a sample, maternal nucleic acid contributes significantly to such mixture, compared to the contribution of fetal alleles. Fetal alleles are identified by their small size. Thus, the relative abundance of fetal nucleic acid in the maternal sample is the total number of unique sequence reads mapped against the target nucleic acid sequence on the reference genome for each of those two alleles at the site of the polymorphism. It can be determined as a parameter.

In certain embodiments, fetal fraction can be determined based on one or more levels. The results of determination of fetal fraction according to level are described, for example, in International Application Publication No. WO2014/055774, the entire contents of which are incorporated herein by reference, including all documents, tables, formulas, and drawings. there is In some embodiments, the fetal fraction is determined according to levels categorized as indicative of maternal and/or fetal copy number variation. For example, determining the fetal fraction can include assessing the expected level of maternal and/or fetal copy number variation that is utilized in determining the fetal fraction. In some embodiments, the fetal fraction is determined for a level (e.g., the first level) categorized as displaying copy number variation according to a range of expected levels determined for the same type of copy number variation. do. The fetal fraction can be determined according to observed levels that fall within expected levels, and are thereby categorized as maternal and/or fetal copy number variations. In some embodiments, the observed level (e.g., the first level) categorizing the fetal fraction as a maternal and/or fetal copy number variant is determined for the same maternal and/or fetal copy number variant. Decide if it is different from the expected level given. Fetal fraction can be provided as a percentage. For example, the fetal fraction can be divided by 100 to give a percentage value. For example, at a first level representing the maternal homozygous duplication, being a level of 155, and an expectation level for the maternal homozygous duplication being a level of 150, the fetal The fraction can be determined as 10% (eg, (fetal fraction = 2 x (155-150)).

The amount of fetal nucleic acid in extracellular nucleic acid can be quantified and used in conjunction with the methods provided herein. Accordingly, in certain embodiments, the methods of the technology described herein comprise the additional step of determining the amount of fetal nucleic acid. The amount of fetal nucleic acid in a nucleic acid sample obtained from a subject can be determined before or after processing to prepare the sample nucleic acid. In certain embodiments, after processing and preparing the sample nucleic acid, the amount of fetal nucleic acid in the sample is determined and this amount is utilized for further evaluation. In some embodiments, the outcome comprises factorizing a fraction of fetal nucleic acid in the sample nucleic acid (e.g., adjusting the count, removing the sample, making a determination, or not making a determination). In certain embodiments, the methods provided herein can be used in conjunction with methods for determining fetal fraction. For example, a method for determining fetal fractions that includes a normalization process can include one or more normalization methods (eg, principal component normalization) provided herein.

The determining step may be included before, during, or at any point in the methods described herein, or in certain particular methods described herein (e.g., detection of aneuploidy, determination of fetal sex). can be done after For example, a method of quantifying fetal nucleic acid is performed before, during, or after determining fetal sex or aneuploidy in order to perform the method of determining fetal sex or aneuploidy with a given sensitivity or specificity. and more than about 2%, 3%, 4%, 5%, 6%, 7%, 8%, 9%, 10%, 11%, 12%, 13%, 14%, 15%, 16%, Samples with 17%, 18%, 19%, 20%, 21%, 22%, 23%, 24%, 25% or more of fetal nucleic acid can be identified. In some embodiments, for example, a sample determined to have a certain threshold amount of fetal nucleic acid (e.g., about 15% or more fetal nucleic acid; about 4% or more fetal nucleic acid) Further analyzes are performed for determination of sex or aneuploidy, or for the presence or absence of aneuploidy or genetic mutations. In certain embodiments, only if the sample has a certain threshold amount of fetal nucleic acid (e.g., about 15% or more fetal nucleic acid; about 4% or more fetal nucleic acid), e.g., fetal nucleic acid Select (e.g., select and communicate to the patient) the gender or presence or absence of aneuploidy determination results.

In some embodiments, determining the fetal fraction or determining the amount of fetal nucleic acid is neither required nor necessary to identify the presence or absence of a chromosomal aneuploidy. In some embodiments, identification of the presence or absence of a chromosomal aneuploidy does not require sequence differentiation between fetal and maternal DNA. In certain embodiments, this is because the combined contribution of both maternal and fetal sequences in a particular chromosome, chromosomal portion or segment thereof is analyzed. In some embodiments, identification of the presence or absence of a chromosomal aneuploidy does not rely on a priori sequence information that would distinguish between fetal and maternal DNA.

Enrichment of Nucleic Acids In some embodiments, nucleic acids (eg, extracellular nucleic acids) are enriched or relatively enriched to obtain subpopulations or species of nucleic acids. Subpopulations of nucleic acids include, for example, fetal nucleic acids, maternal nucleic acids, nucleic acids comprising fragments of a particular length or range of lengths, or a particular genomic region (e.g., a single chromosome, a series of chromosomes and/or a particular (chromosomal regions of the Such enriched samples can be used in conjunction with the methods provided herein. Thus, in certain embodiments, the methods of the present technology include the additional step of enriching for a subpopulation of nucleic acids in the sample, such as fetal nucleic acids. In certain embodiments, the methods described above for determining the fetal fraction can also be used to enrich for fetal nucleic acids. In certain embodiments, maternal nucleic acids are selectively (partially, substantially, almost completely or completely) removed from the sample. In certain embodiments, enrichment to obtain specific low copy number species nucleic acids (eg, fetal nucleic acids) can improve quantitative sensitivity. Methods for enriching a sample for particular species of nucleic acids are described, for example, in U.S. Pat. Publication No. WO2009/032779, International Patent Application Publication No. WO2009/032781, International Patent Application Publication No. WO2010/033639, International Patent Application Publication No. WO2011/034631, International Patent Application Publication No. WO2006/056480 and International Patent Applications Publication No. WO2011/143659, the entire contents of each of which are incorporated herein by reference, including all documents, tables, formulas and drawings.

In some embodiments, nucleic acids are enriched to obtain certain target fragment species and/or reference fragment species. In certain embodiments, nucleic acids are enriched to obtain specific nucleic acid fragment lengths or ranges of fragment lengths using one or more of the methods of length-based separation described below. In certain embodiments, nucleic acids are enriched to select genomic regions using one or more sequence-based separation methods described herein and/or known in the art. A fragment derived from (eg, a chromosome) is obtained. Certain methods for enriching for a subpopulation of nucleic acids in a sample (eg, fetal nucleic acids) are described in detail below.

Some methods for enriching for subpopulations of nucleic acids (e.g., fetal nucleic acids) that can be used with the methods described herein exploit epigenetic differences between maternal and fetal nucleic acids. including how to For example, fetal nucleic acids can be differentiated and separated from maternal nucleic acids based on methylation differences. Methods for enriching fetal nucleic acids based on methylation are described in US Patent Application Publication No. 2010/0105049, which is incorporated herein by reference. Such methods sometimes involve binding a sample nucleic acid to a methylation-specific binding agent (methyl-CpG binding protein (MBD), methylation-specific antibody, etc.) and, based on differences in methylation status, separating bound from unbound nucleic acid. Such methods can also include the use of methylation-sensitive restriction enzymes (described above; e.g., HhaI and HpaII), which selectively and completely or substantially digest maternal nucleic acids to Selective digestion of nucleic acid from a maternal sample with an enzyme that enriches the sample for at least one region of fetal nucleic acid allows for enrichment of the region of fetal nucleic acid in the maternal sample.

Another method for enriching for a subpopulation of nucleic acids (e.g., fetal nucleic acids) that can be used with the methods described herein is U.S. Patent Application Publication No. 2009, which is incorporated herein by reference. Another approach is to enhance polymorphic sequences with restriction endonucleases, such as the method described in US Pat. Such methods include cleaving the nucleic acid containing the non-target allele with a restriction endonuclease that recognizes nucleic acid containing the non-target allele but not the target allele; Amplifying the uncleaved nucleic acid without clotting, wherein the uncleaved, amplified nucleic acid is enriched in target nucleic acid (e.g., fetal nucleic acid) relative to non-target nucleic acid (e.g., maternal nucleic acid). be. In certain embodiments, for example, nucleic acids can be selected to contain alleles with polymorphic sites that are susceptible to selective digestion by a cleaving agent.

Some methods for enriching for subpopulations of nucleic acids (eg, fetal nucleic acids) that can be used with the methods described herein include selective enzymatic degradation approaches. Such methods include protecting target sequences from exonuclease digestion, thereby facilitating elimination of unwanted sequences (eg, maternal DNA) in the sample. For example, in one approach, sample nucleic acid is denatured to produce single-stranded nucleic acids, the single-stranded nucleic acids are contacted with at least one pair of target-specific primers under suitable annealing conditions, and annealed. The resulting primer is extended by polymerization of nucleotides to produce a double-stranded target sequence, and the single-stranded nucleic acid is digested using a nuclease that digests single-stranded (eg, non-target) nucleic acid. In certain embodiments, the method can be repeated for at least one additional cycle. In certain embodiments, the same pair of target-specific primers are used to extend the primers in each of the first and second cycles; For , different, target-specific primer pairs are used.

Some methods for enriching for subpopulations of nucleic acids (eg, fetal nucleic acids) that can be used with the methods described herein include massively parallel signature sequencing (MPSS) approaches. MPSS is a solid-phase method that typically uses ligation of adapters (eg, tags), followed by decoding of the adapters, to read out nucleic acid sequences in aliquots. Typically, the tagged PCR products are amplified, resulting in the production of PCR products with unique tags from each nucleic acid. Tags are often used to tether PCR products to microbeads. After several rounds of ligation-based sequencing, for example, a sequence signature can be identified from each bead. Each signature sequence (MPSS tag) in the MPSS dataset is analyzed and compared to all other signatures and all identical signatures are counted.

In certain embodiments, certain enrichment methods (eg, certain MPS and/or MPSS-based enrichment methods) can include amplification (eg, PCR)-based approaches. In certain embodiments, locus-specific amplification methods can be used (eg, using locus-specific amplification primers). In certain embodiments, a multiplex SNP allele PCR approach can be used. In certain embodiments, a multiplex SNP allele PCR approach can be used in combination with uniplex sequencing. For example, such an approach can involve the use of multiplex PCR (e.g., the MASSARRAY system) and incorporation of capture probe sequences into amplicons followed by sequencing using, for example, the Illumina MPSS system. . In certain embodiments, a multiplex SNP allele PCR approach can be used in combination with a three-primer system and indexed sequencing. For example, such an approach includes a first capture probe incorporated into a forward PCR primer specific to a particular locus, and a locus-specific Using multiplex PCR (e.g., MASSARRAY system) with primers that have adapter sequences incorporated into the reverse PCR primers to generate amplicons, followed by reverse capture sequences and molecular index barcodes. Performing a second PCR for integration can be included. In certain embodiments, a multiplex SNP allele PCR approach can be used in combination with a four primer system and indexed sequencing. For example, such an approach has adapter sequences incorporated into both locus-specific forward and locus-specific reverse PCR primers for sequencing using, for example, the Illumina MPSS system. It can involve using multiplex PCR (e.g., the MASSARRAY system) with primers, followed by a second PCR to incorporate both the forward and reverse capture sequences and the molecular index barcode. . In certain embodiments, microfluidic technology approaches can be used. In certain embodiments, an array-based microfluidic approach can be used. For example, such an approach can involve the use of microfluidic-based arrays (e.g., Fluidigm) for low-plex amplification and incorporation of index and capture probes, followed by sequencing. can. In certain embodiments, for example, emulsion microfluidic technology approaches such as digital droplet PCR can be used.

In certain embodiments, universal amplification methods can be used (eg, using universal primers or non-locus-specific amplification primers). In certain embodiments, universal amplification methods can be used in combination with pull-down approaches. In certain embodiments, the method can comprise a biotinylated ultramer pulldown from a universally amplified sequencing library (eg, a biotinylated pulldown assay from Agilent or IDT). For example, such an approach can involve the steps of standard library preparation, enrichment for selected regions by pull-down assays, and a second universal amplification. In certain embodiments, pull-down approaches can be used in combination with ligation-based methods. In certain embodiments, methods can include pull-down with biotinylated ultramers using ligation of sequence-specific adapters (eg, HALOPLEX PCR, Halo Genomics). For example, such an approach can involve the use of selector probes to capture restriction enzyme digestion fragments, followed by ligation of captured products to adapters and universal amplification, followed by sequencing. In certain embodiments, pull-down approaches can be used in combination with extension and ligation-based methods. In certain embodiments, the method can include extension and ligation with a molecular inversion probe (MIP). For example, such an approach can involve the use of molecular inversion probes in combination with sequence adapters, followed by universal amplification and sequencing. In certain embodiments, complementary DNA can be synthesized and sequenced without amplification.

In certain embodiments, the extension and ligation approach can be performed without a pulldown component. In certain embodiments, the method can include hybridization with locus-specific forward and reverse primers, extension, and ligation. Such methods can further include universal amplification, or complementary DNA synthesis without amplification, followed by sequencing. In certain embodiments, such methods can reduce or eliminate background sequences during analysis.

In certain embodiments, the pull-down approach can be used with or without an optional amplification component. In certain embodiments, methods can include modified pull-down assays and ligations that fully incorporate capture probes and do not perform universal amplification. For example, such approaches include the use of modified selector probes to capture restriction enzyme digested fragments, followed by ligation of captured products to adapters, optional amplification, and sequencing. can be done. In certain embodiments, methods can include biotinylation pull-down assays involving extension and ligation of adapter sequences in combination with circular single-stranded ligation. For example, such approaches can include the use of selector probes to capture regions of interest (e.g., target sequences), probe extension, adapter ligation, single-stranded circular ligation, optional amplification, and sequencing. can. In certain embodiments, analysis of sequencing results can separate target sequences from background.

In some embodiments, nucleic acids are enriched for fragments derived from selected genomic regions (e.g., chromosomes) using one or more of the sequence-based separation methods described herein. obtain. Sequence-based separation generally results in the nucleotide sequence being present in fragments of interest (e.g., target and/or reference fragments) and substantially absent in other fragments of the sample, or insignificant in other fragments. Based on being present in only minor amounts (eg 5% or less). In some embodiments, sequence-based separation can involve separation of target fragments and/or separation of reference fragments. The separated target fragment and/or the separated reference fragment are often isolated and removed from the remaining fragments in the nucleic acid sample. In certain embodiments, the separated target fragment and the separated reference fragment are also isolated from each other and removed (eg, isolated as compartments in a separation assay). In certain embodiments, the separated target fragment and the separated reference fragment are isolated together (eg, isolated as the same assay compartment). In some embodiments, unbound fragments can be differentially removed or degraded or digested.

In some embodiments, a process that selectively captures nucleic acids is used to separate and remove target and/or reference fragments from a nucleic acid sample. Commercially available systems for capturing nucleic acids include, for example, the Nimblegen Sequence Capture System (Roche NimbleGen, Madison, Wis.); the Illumina BEADARRAY platform (Illumina, San Diego, CA); the Affymetrix GENECHIP platform (Affymetrix, Santa Clara, CA). Agilent SureSelect Target Enrichment System (Agilent Technologies, Santa Clara, Calif.); and related platforms. Such methods typically involve hybridization of a capture oligonucleotide to a segment or all of the nucleotide sequence of a target fragment or reference fragment, and use of solid phase (e.g., solid phase arrays) and/or solution-based platforms. can include capture oligonucleotides so as to preferentially hybridize to nucleic acid fragments from a selected genomic region or locus (e.g., one of the 21st, 18th, 13th, X or Y chromosomes, or the reference chromosome) (sometimes called "decoys"). In certain embodiments, hybridization-based methods (e.g., using oligonucleotide arrays) are used to enrich for certain chromosomes (e.g., potential aneuploid chromosomes, reference chromosomes) , or other chromosome of interest), or a segment thereof of interest, can be obtained.

In some embodiments, one or more methods of length-based separation are used to separate nucleic acids from particular nucleic acid fragment lengths, range lengths, or below a particular threshold or cutoff. Or enrich for greater lengths. The length of nucleic acid fragments typically refers to the number of nucleotides in the fragment. The length of a nucleic acid fragment is also sometimes referred to as the size of the nucleic acid fragment. In some embodiments, methods of length-based separation are performed without measuring the length of individual fragments. In some embodiments, methods of length-based separation are performed in conjunction with methods for determining the length of individual fragments. In some embodiments, length-based separation refers to size fractionation procedures, and all or a portion of the fractionated pool can be isolated (eg, retained) and/or analyzed. Size fractionation procedures are known in the art (e.g., separation on arrays, separation by molecular sieves, separation by gel electrophoresis, separation by column chromatography (e.g., molecular sieve columns), and microfluidic separations). technology-based approach). In certain embodiments, length-based separation approaches include, for example, fragment circularization, chemical treatment (e.g., formaldehyde, polyethylene glycol (PEG)), mass spectrometry, and/or size-specific Mention may be made of nucleic acid amplification.

Certain length-based separation methods that can be used with the methods described herein utilize, for example, selective sequence tagging approaches. The term "sequence tagging" refers to the incorporation of recognizable and distinct sequences into a nucleic acid or population of nucleic acids. The term "sequence tagging" as used herein has a different meaning than the term "sequence tag" described later in this specification. In such a sequence tagging method, nucleic acids of certain fragment size species (eg, short fragments) are subjected to selective sequence tagging in samples containing long and short nucleic acids. Such methods typically involve performing a nucleic acid amplification reaction using a set of nested primers comprising an inner primer and an outer primer. In certain embodiments, one or both of the inner primers may be tagged, thereby introducing the tag onto the target amplicon. The outer primer generally does not anneal to the short fragment carrying the (inner) target sequence. The inner primer can anneal to the short fragment to generate an amplification product carrying the tag and target sequence. Typically, long fragment tagging is inhibited through a combination of mechanisms including, for example, blocking extension of the inner primer by previous annealing and extension of the outer primer. Tagging by any of a variety of methods including, for example, exonuclease digestion of the single-stranded nucleic acid and amplification of the tagged fragment using at least one tag-specific amplification primer. Enrichment can be performed on the fragments obtained.

Another method of length-based separation that can be used with the methods described herein involves subjecting a nucleic acid sample to polyethylene glycol (PEG) precipitation. Examples of methods include those described in International Patent Application Publication Nos. WO2007/140417 and WO2010/115016, the entire contents of each of which is incorporated herein by reference, including all documents, tables, formulas, and drawings. incorporated herein by The method generally involves the addition of one or more monovalent salts under conditions sufficient to substantially precipitate large nucleic acids without substantially precipitating small (e.g., less than 300 nucleotides) nucleic acids. In its presence, it requires contacting the nucleic acid sample with PEG.

Another size-based enrichment method that can be used with the methods described herein involves circularization by ligation, eg, ligation using circligase. Short nucleic acid fragments can typically be circularized with greater efficiency than longer fragments. The non-circularized sequences can be separated from the circularized sequences and the enriched short fragments can be used for further analysis.

Nucleic Acid Libraries In some embodiments, a nucleic acid library is prepared by a specific treatment, including non-limiting examples thereof, immobilization on a solid phase (e.g., solid support, e.g., flow cell, beads), A plurality of polynucleotide molecules (e.g., a sample of nucleic acids) that are prepared, assembled, and/or modified for nucleic acid sequencing, including enrichment, amplification, cloning, detection) and/or for nucleic acid sequencing. is. In certain embodiments, a nucleic acid library is prepared prior to or during the sequencing process. Nucleic acid libraries (eg, sequencing libraries) can be prepared by any suitable method known in the art. Nucleic acid libraries can be prepared by targeted or non-targeted preparation processes.

In some embodiments, a library of nucleic acids is modified to include chemical moieties (eg, functional groups) configured for immobilization of nucleic acids to solid supports. In some embodiments, a library of nucleic acids is modified to contain biological molecules (e.g., functional groups) and/or binding pairs configured for immobilization of the library to a solid support. Non-limiting examples thereof, including members, thyroxine-binding globulin, steroid-binding protein, antibody, antigen, hapten, enzyme, lectin, nucleic acid, repressor, protein A, protein G, avidin, streptavidin, biotin , complement component C1q, nucleic acid binding proteins, receptors, carbohydrates, oligonucleotides, polynucleotides, complementary nucleic acid sequences, etc., and combinations thereof. Some examples of specific binding pairs include, but are not limited to, avidin and biotin moieties; antigenic epitopes and antibodies or immunologically reactive fragments thereof; antibodies and haptens; fluorescein moieties and anti-fluorescein antibodies; operators and repressors; nucleases and nucleotides; lectins and polysaccharides; steroids and steroid-binding proteins; and substrates; immunoglobulins and Protein A; oligonucleotides or polynucleotides and their corresponding complements, etc., or combinations thereof.

In some embodiments, a library of nucleic acids is modified to contain one or more polynucleotides of known composition, non-limiting examples of which are identifiers (e.g., tags, index tags), capture Sequences, labels, adapters, restriction enzyme sites, promoters, enhancers, origins of replication, stem loops, complementary sequences (e.g. primer binding sites, annealing sites), suitable integration sites (e.g. transposons, viral integration sites), modified nucleotides etc., or combinations thereof. Polynucleotides of known sequence can be added at appropriate positions, eg, at the 5' end, 3' end or internally of the nucleic acid sequence. Polynucleotides of known sequence may be of the same sequence or of different sequences. In some embodiments, a polynucleotide of known sequence is configured to hybridize to one or more oligonucleotides immobilized on a surface (eg, a surface in a flow cell). For example, a nucleic acid molecule containing a 5' known sequence can be hybridized to a first, plurality of oligonucleotides, while the 3' known sequence of the molecule is hybridized to a second, plurality of oligonucleotides. can be made In some embodiments, the library of nucleic acids can include chromosome-specific tags, capture sequences, labels and/or adapters. In some embodiments, the library of nucleic acids comprises one or more detectable labels. In some embodiments, one or more detectable labels are placed in the nucleic acid library at the 5′ end, at the 3′ end, and/or at any nucleotide position within the nucleic acids in the library. can be incorporated in In some embodiments, the library of nucleic acids comprises hybridized oligonucleotides. In certain embodiments, the hybridized oligonucleotides are labeled probes. In some embodiments, the library of nucleic acids comprises oligonucleotide probes hybridized prior to immobilization on the solid phase.

In some embodiments, a polynucleotide of known sequence comprises a universal sequence. A universal sequence is a specific nucleotide sequence that incorporates into two or more nucleic acid molecules, or two or more subsets of nucleic acid molecules; The same is true for all molecules. Universal sequences are often designed to hybridize to and/or amplify multiple different sequences using a single universal primer complementary to the universal sequence. In some embodiments, two (eg, pairs) or more universal sequences and/or universal primers are used. Universal primers often contain a universal sequence. In some embodiments, an adapter (eg, universal adapter) comprises a universal sequence. In some embodiments, one or more universal sequences are used to capture, identify and/or detect multiple species or subsets of nucleic acids.

In certain embodiments of nucleic acid library preparation (e.g., in the case of certain sequencing by synthetic procedures), nucleic acids are selected and/or fragmented by size to a few hundred base pairs or more. (eg, in preparation for library generation). In some embodiments, library preparation is performed without fragmentation (eg, when using ccfDNA).

In certain embodiments, ligation-based library preparation methods are used (eg, ILLUMINA TRUSEQ, Illumina, San Diego Calif.). Ligation-based library preparation methods often exploit the design of adapters (e.g., methylated adapters) that can incorporate index sequences in the initial ligation step, often from a single end. It can be used to prepare samples for read sequencing, double-end sequencing, and multiplex sequencing. End repair of nucleic acids (eg, fragmented nucleic acids or ccfDNA) is sometimes performed, for example, by fill-in reactions, exonuclease reactions, or combinations thereof. In some embodiments, the resulting blunt end repair nucleic acid can then be extended by a single nucleotide that is complementary to the single nucleotide overhang on the 3' end of the adapter/primer. Any nucleotide can be used for the extension/overhang nucleotide. In some embodiments, preparation of the nucleic acid library comprises ligation of adapter oligonucleotides. Adapter oligonucleotides are often complementary to flow cell anchors and are sometimes utilized, for example, to immobilize nucleic acid libraries to a solid support, such as the inner surface of a flow cell. In some embodiments, the adapter oligonucleotide comprises an identifier, one or more sequencing primer hybridization sites (e.g., a sequence exhibiting complementarity to a universal sequencing primer, a single end sequencing primer, both ends sequencing primers, multiplexed sequencing primers, etc.), or combinations thereof (eg, adapter/sequencing, adapter/identifier, adapter/identifier/sequencing).

An identifier is a suitable detectable label incorporated into or tethered to a nucleic acid (eg, a polynucleotide) that allows detection and/or identification of the nucleic acid containing it. In some embodiments, the identifier is incorporated into or tethered to the nucleic acid (eg, by a polymerase) during sequencing. Non-limiting examples of identifiers include nucleic acid tags, nucleic acid indexes or barcodes, radiolabels (e.g., isotopes), metal labels, fluorescent labels, chemiluminescent labels, phosphorescent labels, fluorophore quenchers, dyes, proteins (eg, enzymes, antibodies or portions thereof, linkers, members of binding pairs), etc., or combinations thereof. In some embodiments, an identifier (eg, a nucleic acid index or barcode) is a nucleotide or nucleotide analogue of a unique, known and/or identifiable sequence. In some embodiments, the identifier is 6 or more contiguous nucleotides. A large number of fluorophores are available with a variety of different excitation and emission spectra. Any suitable type and/or number of fluorophores can be used as identifiers. In some embodiments, one or more, two or more, three or more, four or more, five or more, six or more, seven or more, eight Or more, 9 or more, 10 or more, 20 or more, 30 or more, or 50 or more different identifiers can be used in the methods described herein (e.g., nucleic acid detection and/or sequencing methods). In some embodiments, one or two types of identifiers (eg, fluorescent labels) are linked to each nucleic acid in the library. Detection and/or quantification of identifiers can be performed by any suitable method, machine or apparatus, non-limiting examples of which include flow cytometry, quantitative polymerase chain reaction (qPCR), gel electrophoresis, luminescence. meter, fluorometer, spectrophotometer, analysis by suitable gene chip or microarray, western blot, mass spectrometry, chromatography, analysis by cytofluorometry, fluorescence microscopy, suitable fluorescence or digital imaging, confocal Laser scanning microscopy, laser scanning cytometry, affinity chromatography, manual batch mode separation, electric field suspension, suitable nucleic acid sequencing methods and/or nucleic acid sequencing instruments, etc., and combinations thereof.

In some embodiments, transposon-based library preparation methods are used (eg, EPICENTRE NEXTERA, Epicentre, Madison Wis.). Transposon-based methods typically use in vitro transposition to simultaneously fragment and tag DNA (often with platform-specific tags and optional barcodes can be incorporated) to prepare a library that can be used on a sequencer.

In some embodiments, the nucleic acid library or portion thereof is amplified (eg, amplified by PCR-based methods). In some embodiments, the sequencing method comprises amplification of a nucleic acid library. A nucleic acid library can be amplified before or after immobilization on a solid support (eg, a solid support in a flow cell). Nucleic acid amplification amplifies or increases the number of nucleic acid templates and/or their complements present (e.g., in a nucleic acid library) by generating one or more copies of the templates and/or their complements. It includes processing to cause Amplification can be performed by any suitable method. Nucleic acid libraries can be amplified by thermocycling or isothermal amplification methods. In some embodiments, rolling circle amplification methods are used. In some embodiments, amplification occurs on a solid support (eg, inside a flow cell) to which the nucleic acid library or portion thereof is immobilized. In one particular sequencing method, a nucleic acid library is added to a flow cell and immobilized by hybridization to anchors under appropriate conditions. This type of nucleic acid amplification is often referred to as solid phase amplification. In some embodiments of solid-phase amplification, all or part of the amplification product is synthesized by extension starting from immobilized primers. Solid-phase amplification reactions are similar to standard solution-phase amplification, except that at least one of the amplification oligonucleotides (eg, primers) is immobilized on a solid support.

In some embodiments, solid phase amplification comprises a nucleic acid amplification reaction comprising only one species of oligonucleotide primer immobilized on a surface. In certain embodiments, solid phase amplification comprises multiple different immobilized oligonucleotide primer species. In some embodiments, solid phase amplification can comprise a nucleic acid amplification reaction comprising one species of oligonucleotide primer immobilized on a solid surface and a second, different oligonucleotide primer species in solution. . Multiple different species of immobilized primers or solution-based primers can be used. Non-limiting examples of solid-phase nucleic acid amplification reactions include interface amplification, bridge amplification, emulsion PCR, WildFire amplification (eg, US Patent Publication No. US20130012399), etc., or combinations thereof.

Sequencing In some embodiments, nucleic acids (eg, nucleic acid fragments, sample nucleic acids, cell-free nucleic acids) are sequenced. In certain embodiments, complete or substantially complete sequences are obtained, and sometimes partial sequences are obtained.

In some embodiments, some or all of the nucleic acids in the sample are enriched and/or amplified (eg, non-specifically, eg, by PCR-based methods) prior to or during sequencing. In certain embodiments, specific portions or subsets of nucleic acids in a sample are enriched and/or amplified prior to or during sequencing. In some embodiments, sequencing of a portion or subset of a preselected pool of nucleic acids is performed randomly. In some embodiments, nucleic acids in the sample are not enriched and/or amplified prior to or during sequencing.

As used herein, "reads" (e.g., "a read", "a sequence read") refer to any sequence described herein or known in the art. is a short nucleotide sequence generated by any sequencing process. Reads can be generated from one end of a nucleic acid fragment ("reads from a single end"), and sometimes from both ends of a nucleic acid (e.g., reads from both ends, reads from two ends). lead).

Sequence read length is often associated with a particular sequencing technique. For example, high-throughput methods provide sequence reads that can vary in size from tens to hundreds of base pairs (bp). For example, nanopore sequencing can provide sequence reads that can vary in size from tens to hundreds or thousands of base pairs. In some embodiments, the mean, median, mean or absolute length of the sequence reads is from about 15 bp to about 900 bp in length. In certain embodiments, the mean, median, mean or absolute length of the sequence reads is about 1000 bp or greater.

In some embodiments, the nominal average value, average length or absolute length of reads from a single end is sometimes from about 15 contiguous nucleotides to about 50 or more contiguous nucleotides. from about 40 contiguous nucleotides to about 40 or more contiguous nucleotides, sometimes about 15 contiguous nucleotides, or about 36 or more contiguous nucleotides. In certain embodiments, the nominal average value, average length, or absolute length of reads from a single end is about 20 to about 30 bases long, or about 24 to about 28 bases long. In certain embodiments, the nominal average value, average length or absolute length of reads from a single end is about 1, about 2, about 3, about 4, about 5, about 6, about 7, about 8, about 9, about 10, about 11, about 12, about 13, about 14, about 15, about 16, about 17, about 18, about 19, about 21, about 22, about 23, about 24, about 25 , about 26, about 27, about 28, or about 29 bases long or longer.

In certain embodiments, the nominal average value, average length or absolute length of reads from both ends sometimes ranges from about 10 contiguous nucleotides to about 25 contiguous nucleotides or more (e.g., about 10, about 11, about 12, about 13, about 14, about 15, about 16, about 17, about 18, about 19, about 20, about 21, about 22, about 23, about 24 or about 25 nucleotides long or less greater than), from about 15 contiguous nucleotides to about 20 contiguous nucleotides or more, sometimes about 17 contiguous nucleotides, or about 18 contiguous nucleotides.

A read is generally a physical nucleic acid representation of a nucleotide sequence. For example, in a read containing a sequence depicted as ATGC, "A" designates an adenine nucleotide, "T" designates a thymine nucleotide, "G" designates a guanine nucleotide, as physical nucleic acids; "C" denotes a cytosine nucleotide. Sequence reads obtained from pregnant female blood may be reads derived from a mixture of fetal and maternal nucleic acids. A mixture of relatively short reads can be converted by the processing described herein into representations of genomic nucleic acid present in pregnant females and/or fetuses. A mixture of relatively short reads can be converted, for example, to display copy number variation (eg, maternal and/or fetal copy number variation), genetic variation, or aneuploidy. A mixture of maternal and fetal nucleic acid reads can be transformed into a representation of a composite chromosome or segments thereof containing features of one or both of the maternal and fetal chromosomes. In certain embodiments, "obtaining" nucleic acid sequence reads of a sample obtained from a subject and/or "obtaining" nucleic acid sequence reads of a biological specimen obtained from one or more reference persons ” can include directly sequencing the nucleic acid to obtain the sequence information. In some embodiments, "obtaining" can include receiving sequence information obtained directly from the nucleic acid by another.

In some embodiments, the represented proportion of the genome is sequenced, sometimes referred to as "coverage" or "fold coverage." For example, 1× coverage indicates that approximately 100% of the nucleotide sequence of the genome is represented by the reads. In some embodiments, "fold coverage" is a term that is compared with reference to a previous sequencing run as a reference. For example, the second sequencing run may be half the coverage of the first sequencing run. In some embodiments, the genome is sequenced with redundancy, where a given region of the genome can be covered by two or more reads, or overlapping reads. (eg, a “coverage factor” greater than 1, eg, 2× coverage).

In some embodiments, one nucleic acid sample obtained from one individual is sequenced. In certain embodiments, nucleic acids obtained from each of two or more samples are sequenced, where the samples are obtained from one individual or from different individuals. In certain embodiments, nucleic acid samples obtained from two or more biological samples are pooled, where each biological sample represents one individual, or two or more individuals. Sequencing of pooled samples obtained from . In the latter embodiment, nucleic acid samples obtained from each biological sample are often identified by one or more unique identifiers.

In some embodiments, the sequencing method utilizes identifiers that allow multiplexing of sequencing reactions in the sequencing process. The greater the number of unique identifiers, the greater the number of detected samples and/or chromosomes that can be multiplexed, eg, in a sequencing process. Any suitable number (eg, 4, 8, 12, 24, 48, 96 or more) of unique identifiers can be used for the sequencing process.

Sequencing processes sometimes employ a solid phase, sometimes comprising a flow cell, onto which nucleic acids from a library can be tethered, reagents being flowed into contact with the tethered nucleic acids. can be done. A flow cell sometimes includes lanes of the flow cell, and the use of identifiers can facilitate analysis of several samples in each lane. A flow cell is often a solid support that can be configured to hold bound analytes and/or allow the orderly passage of reagent solutions over bound analytes. Flow cells often have a planar geometry, are optically transparent, are generally of millimeter or sub-millimeter scale, and often have channels or lanes in which analytes and reagents are stored. interaction occurs. In some embodiments, the number of samples analyzed in a given lane of the flow cell depends on the number of unique identifiers utilized during library preparation and/or probe design. Lane of a single flow cell. For example, multiplexing using 12 identifiers allows simultaneous analysis of 96 samples in an 8-lane flow cell (e.g., equal to the number of wells in a 96-well microwell plate). . Similarly, analyzing 384 samples simultaneously in an 8-lane flow cell (e.g., equal to the number of wells in a 384-well microwell plate) by multiplexing, e.g., using 48 identifiers. be possible. Non-limiting examples of commercially available multiplexed sequencing kits include Illumina's Multiplexed Sample Preparation Oligonucleotide Kit, and Multiplexed Sequencing Primer and PhiX Control Kits (e.g., Illumina's Cat. 400-1001 and PE-400-1002).

Any suitable method of sequencing nucleic acids can be used, non-limiting examples of which are Maxim & Gilbert, chain termination, sequencing by synthesis, sequencing by ligation, sequencing by mass spectrometry. , microscopy-based techniques, etc., or combinations thereof. In some embodiments, the methods provided herein use first generation technologies, such as Sanger sequencing, which include automated Sanger sequencing, including microfluidic Sanger sequencing. ) can be used. In some embodiments, sequencing techniques can be used, including the use of nucleic acid imaging techniques (eg, transmission electron microscopy (TEM) and atomic force microscopy (AFM)). In some embodiments, high throughput sequencing methods are used. High-throughput sequencing methods generally involve clonally amplifying DNA templates or single DNA molecules, and sequencing these templates or molecules in massively parallel, sometimes inside flow cells. Next generation (e.g., second and third generation) sequencing technologies capable of massively parallel DNA sequencing can be used for the methods described herein. are collectively referred to herein as "massively parallel sequencing" (MPS). In some embodiments, MPS sequencing utilizes a targeted approach, where a specific chromosome, gene, or region of interest is sequenced. In certain embodiments, a non-targeted approach is used, in which most or all nucleic acids in a sample are sequenced, amplified, and/or captured at random.

In some embodiments, targeted approaches of enrichment, amplification and/or sequencing are used. Targeted approaches often isolate, select and/or enrich a subset of nucleic acids in a sample for further processing through the use of sequence-specific oligonucleotides. In some embodiments, a library of sequence-specific oligonucleotides is utilized to target (eg, hybridize to) one or more sets of nucleic acids in a sample. Sequence-specific oligonucleotides and/or primers often select specific sequences (e.g., unique nucleic acid sequences) present in one or more of the chromosomes, genes, exons, introns and/or regulatory regions of interest Let Any suitable method or combination of methods can be used to enrich, amplify and/or sequence one or more subsets of targeted nucleic acids. In some embodiments, targeted sequences are isolated and/or enriched by capture onto a solid phase (eg, flow cell, bead) using one or more sequence-specific anchors. In some embodiments, targeted sequences are identified by polymerase-based methods (e.g., PCR-based methods with any suitable polymerase-based extension) using sequence-specific primers and/or primer sets. Enrich and/or amplify. Sequence-specific anchors can often be used as sequence-specific primers.

MPS sequencing sometimes uses sequencing-by-synthesis and certain visualization processes. Nucleic acid sequencing techniques that can be used in the methods described herein include sequencing-by-synthesis and reversible chain-terminating nucleotide-based sequencing (e.g., Illumina's Genome Analyzer; Genome Analyzer II; HISEQ2000; HISEQ2500 ( Illumina, San Diego Calif.)). Using this technology, millions of nucleic acid (eg, DNA) fragments can be sequenced in parallel. One example of this type of sequencing technology uses a flow cell containing optically transparent slides with eight individual lanes, onto whose surface oligonucleotide anchors (e.g., adapter primers) are attached. is doing. A flow cell is often a solid support that can be configured to hold bound analytes and/or allow the orderly passage of reagent solutions over bound analytes. Flow cells often have a planar geometry, are optically transparent, are generally of millimeter or sub-millimeter scale, and often have channels or lanes in which analytes and reagents are stored. interaction occurs.

In some embodiments, sequencing-by-synthesis involves repetitive (eg, covalent addition) addition of nucleotides to a primer or an existing nucleic acid strand in a template-directed manner. Each time a nucleotide is added, detection is performed and the process is repeated multiple times until the sequence of the nucleic acid strand is obtained. The length of the resulting sequence will depend, in part, on the number of addition and detection steps performed. In some embodiments of sequencing by synthesis, 1, 2, 3 or more nucleotides of the same type (e.g., A, G, C or T) are added and detected in a single nucleotide addition. do. Nucleotides can be added by any suitable (eg, enzymatic or chemical) method. For example, in some embodiments, a polymerase or ligase is directed to a template to add nucleotides to a primer or an existing nucleic acid strand. Some embodiments of sequencing-by-synthesis use different types of nucleotides, nucleotide analogs and/or identifiers. In some embodiments, reversible chain-terminating nucleotides and/or removable (eg, cleavable) identifiers are used. In some embodiments, fluorescently labeled nucleotides and/or nucleotide analogs are used. In certain embodiments, sequencing-by-synthesis includes cleavage (eg, cleavage and removal of identifiers) and/or washing steps. In some embodiments, addition of one or more nucleotides is detected by any suitable method described herein or known in the art, and non-limiting examples thereof include any suitable suitable cameras, digital cameras, imagers based on CCD (charge coupling device) (e.g. CCD cameras), imagers based on CMOS (Complementary Metal Oxide Silicon) (e.g. , CMOS cameras), photodiodes (e.g., photomultiplier tubes), electron microscopy, field effect transistors (e.g., DNA field effect transistors), ISFET ion sensors (e.g., CHEMFET sensors), etc., or combinations thereof. . Other sequencing methods that can be used to practice the methods herein include sequencing by digital PCR and hybridization.

Other sequencing methods that can be used to practice the methods herein include sequencing by digital PCR and hybridization. Digital polymerase chain reaction (digital PCR or dPCR) can be used to directly identify and quantify nucleic acids in a sample. In some embodiments, digital PCR can be performed in emulsion. For example, individual nucleic acids are separated, eg, in a microfluidic chamber device, and each nucleic acid is individually amplified by PCR. Nucleic acids can be separated such that only one nucleic acid is present per well. In some embodiments, different probes can be used to distinguish between different alleles (eg, fetal and maternal alleles). Alleles can be enumerated to determine copy number.

In certain embodiments, sequencing by hybridization can be used. The method includes contacting a plurality of polynucleotide sequences with a plurality of polynucleotide probes, each of the plurality of polynucleotide probes optionally tethered to a substrate. In some embodiments, the substrate can be a flat surface bearing an array of known nucleotide sequences. The pattern of hybridization to the array can be used to determine the polynucleotide sequences present in the sample. In some embodiments, each probe is tethered to a bead, such as a magnetic bead. Identifies hybridization to beads and can be used to identify multiple polynucleotide sequences within a sample.

In some embodiments, nanopore sequencing can be used in the methods described herein. Nanopore sequencing is a single-molecule sequencing technology whereby the sequence of a single nucleic acid molecule (eg, DNA) is directly determined each time it passes through a nanopore.

Nucleic acid sequence reads can be obtained using any MPS method, system or technology platform suitable for the methods of practice described herein. Non-limiting examples of MPS platforms include Illumina/Solex/HiSeq (e.g., Illumina's Genome Analyzer; Genome Analyzer II; HISEQ2000; HISEQ), SOLiD, Roche/454, PACBIO and/or SMRT, Helicos True Single Molecule Sequence, Sequencing based on Ion Torrent and ionic semiconductors (e.g. developed by Life Technologies), WildFire, 5500, 5500xl W and/or 5500xl W Genetic Analyzer based technologies (e.g. developed and marketed by Life Technologies, USA Patent Publication No. US20130012399); Polony Sequencing, Pyrosequencing, Massively Parallel Signature Sequencing (MPSS), RNA Polymerase (RNAP) Sequencing, LaserGen Systems and Methods, Nanopore-based Platforms, Chemisensitive Field Effect Transistor (CHEMFET) arrays, electron microscopy-based sequencing (eg, ZS Genetics, developed by Halcyon Molecular), nanoball sequencing, etc., or combinations thereof.

In some embodiments, chromosome-specific sequencing is performed. In some embodiments, DANSR (Digital Analysis of Selected Regions) is utilized to perform chromosome-specific sequencing. Hundreds of loci were analyzed by digital analysis of selected regions by cfDNA-dependent catenation of two locus-specific oligonucleotides via intervening "bridge" oligonucleotides to form PCR templates. can be quantified at the same time. In some embodiments, chromosome-specific sequencing is performed by generating a library enriched for chromosome-specific sequences. In some embodiments, sequence reads are obtained only for a selected set of chromosomes. In some embodiments, sequence reads are obtained for chromosomes 21, 18 and 13 only. In some embodiments, sequence reads are obtained and/or mapped against the entire reference genome or a segment of the genome.

In some embodiments, sequence reads are generated, obtained, collected, accumulated, manipulated, transformed, processed, and/or provided by a sequence module. A machine comprising a sequence module can be any suitable machine and/or apparatus for sequencing nucleic acids utilizing sequencing techniques known in the art. In some embodiments, the sequencing module can align, assemble, fragment, complement, reverse complement, and/or error-check (error-correct sequence reads).

In some embodiments, the nucleotide sequence reads obtained from the sample are partial nucleotide sequence reads. As used herein, a "partial nucleotide sequence read" refers to a sequence read of any length with incomplete sequence information, also referred to as sequence ambiguity. A partial nucleotide sequence read may lack information regarding nucleobase identity and/or nucleobase position or order. Partial nucleotide sequence reads generally include sequence reads where only incomplete sequence information (or less than all of the bases have been sequenced or determined) resulted from inadvertent or unintentional sequencing errors. do not have. Such sequencing errors may be inherent in a particular sequencing process and may include, for example, imprecise determination of nucleobase identity and missing or extra nucleobases. Therefore, for partial nucleotide sequence reads herein, certain information about the sequence is often deliberately excluded. That is, sequence information concerning less than all of the nucleobases, or which may or may otherwise be characterized as sequencing errors, is deliberately obtained. In some embodiments, the partial nucleotide sequence read may span a portion of the nucleic acid fragment. In some embodiments, the partial nucleotide sequence read may span the entire length of the nucleic acid fragment. Partial nucleotide sequence reads are described, for example, in International Patent Application Publication No. WO2013/052907, the entire contents of which are incorporated herein by reference, including all documents, tables, formulas, and figures. ing.

Mapping of Reads The reads of a sequence can be mapped. Any suitable method of mapping (eg, processes, algorithms, programs, software, modules, etc., or combinations thereof) can be used, and certain aspects of the mapping process are described below.

Mapping of nucleotide sequence reads (e.g., sequence information obtained from fragments whose physical location in the genome is unknown) can be performed in several ways, often by mapping the resulting sequence reads. , with matching sequences in the reference genome. In such an alignment, the sequence reads are generally aligned to a reference sequence and the aligned reads are referred to as being "mapped", "mapped sequence reads" or "mapped reads". .

As used herein, the terms "aligned," "alignment," or "aligning" can identify a match (e.g., 100% identity) or a partial match. Reference is made to one or more nucleic acid sequences. Alignments can be performed manually or by computer (e.g., software, programs, modules or algorithms), non-limiting examples of which are Efficient Local Alignment distributed as part of the Illumina Genomics Analysis pipeline. of Nucleotide Data (ELAND) computer program. Alignment of sequence reads can be 100% sequence identity. In some cases, the alignment is less than 100% sequence match (eg, imperfect match, partial match, partial alignment). In some embodiments, the alignment is about 99%, 98%, 97%, 96%, 95%, 94%, 93%, 92%, 91%, 90%, 89%, 88%, 87%, 86%, 85%, 84%, 83%, 82%, 81%, 80%, 79%, 78%, 77%, 76% or 75% concordance. In some embodiments, alignments contain mismatches. In some embodiments the alignment contains 1, 2, 3, 4 or 5 mismatches. Two or more sequences can be aligned using either strand. In certain embodiments, a nucleic acid sequence is aligned with the reverse complement of another nucleic acid sequence.

A variety of computational methods can be used to map and/or align sequence reads to a reference genome. Non-limiting examples of computer algorithms that can be used to align sequences include BLAST, BLITZ, FASTA, BOWTIE1, BOWTIE2, ELAND, MAQ, PROBEMATCH, SOAP or SEQMAP, or variations or combinations thereof. include, but are not limited to: In some embodiments, sequence reads can be aligned with sequences in a reference sequence and/or reference genome. In some embodiments, the sequence reads are stored in nucleic acid databases known in the art, including, for example, GenBank, dbEST, dbSTS, EMBL (European Molecular Biology Laboratory) and DDBJ (DNA Databank of Japan). Headings and/or sequences within them may be aligned. The identified sequences can be searched against sequence databases using BLAST or similar tools.

In some embodiments, the mapped sequence reads and/or information associated with the mapped sequence reads are stored on a non-transitory computer readable storage medium in a suitable computer readable format and/or accessed from there. As used herein, "computer-readable format" is sometimes loosely referred to as format. In some embodiments, mapped sequence reads are stored and/or accessed in a suitable binary format, text format, etc., or a combination thereof. The binary format is sometimes the BAM format. The text format is sometimes a sequence alignment/map (SAM) format. Non-limiting examples of binary and/or text formats include BAM, SAM, SRF, FASTQ, Gzip, etc., or combinations thereof. In some embodiments, the mapped sequence reads are stored in a format that requires less storage space (e.g., fewer bytes) than conventional formats (e.g., SAM or BAM formats), and/ or converted to it. In some embodiments, the mapped sequence reads in the first format are compressed into a second format that requires less storage space than the first format. The term "compressed," as used herein, refers to processes of data compression, source encoding, and/or bitrate reduction that reduce the size of computer-readable data files. In some embodiments, the mapped sequence reads are compressed from the SAM format in binary format. When compressing files, some data is sometimes lost. Sometimes no data is lost in the compression process. In some embodiments of file compression, some data is replaced with an index and/or reference to another data file containing information about the reads of the mapped sequences. In some embodiments, the reads of the mapped sequences are represented by a read count, a chromosomal identifier (e.g., identifying the chromosome to which the read maps), and a chromosomal position (e.g., to which the read maps). are stored in a binary format containing or consisting of identifiers of chromosomal locations. In some embodiments, the binary format includes a 20-byte array, a 16-byte array, an 8-byte array, a 4-byte array, or a 2-byte array. In some embodiments, the mapped read information is in 10-byte format, 9-byte format, 8-byte format, 7-byte format, 6-byte format, 5-byte format, 4-byte format, 3-byte format or 2-byte format. Store in an array. Sometimes mapped read data is stored in a 4-byte array containing a 5-byte format. In some embodiments, the binary format includes a 5-byte format that includes a 1-byte chromosomal ordinal number and a 4-byte chromosomal position. In some embodiments, the mapped reads are about 1/100, about 1/90, about 1/80, about 1/70, about 1/60, about 1/1 of the sequence alignment/map (SAM) format. 55, about 1/50, about 1/45, about 1/40, or about 1/30, in a compressed binary format. In some embodiments, the mapped reads are about 1/2 to about 1/50 of the GZip format (eg, about 1/30, 1/25, 1/20, 1/19, 1/18, 1 /17, 1/16, 1/15, 1/14, 1/13, 1/12, 1/11, 1/10, 1/9, 1/8, 1/7, 1/6 or about 1/ 5) is stored in a compressed binary format.

In some embodiments, the system includes a compression module (eg, 4 in FIG. 10A). In some embodiments, a compression module compresses mapped sequence read information stored on a non-transitory computer-readable storage medium in a computer-readable format. The compression module sometimes converts the mapped sequence reads to and from the appropriate format. In some embodiments, the compression module receives mapped sequence reads in a first format (e.g., 1), converts them to a compressed format (e.g., binary format, 5), compresses The leads can be transferred to another module (eg, biased density module, 6). Compression modules often provide sequence reads in binary format, 5 (eg, BReads format). Non-limiting examples of compression modules include GZIP, BGZF and BAM, etc., or variations thereof.
Below is an example of converting an integer to a 4-byte array using java.

In some embodiments, reads can be mapped uniquely or non-uniquely to a reference genome. A read is considered "uniquely mapped" if it aligns to a single sequence in the reference genome. A read is considered "non-uniquely mapped" if it aligns with two or more sequences in the reference genome. In some embodiments, non-uniquely mapped reads are excluded from further analysis (eg, quantification). In certain embodiments, specific, low degree discrepancies (0-1) may exist between the reference genome and the reads from individual samples being mapped. can be explained as In some embodiments, reads mapped against a reference sequence are not allowed to have any degree of mismatch.

As used herein, the term "reference genome" is any particular known sequenced or characterized genome of any organism or virus, whether partial or complete. , can refer to a genome that can be used to refer to identified sequences from a subject. A reference genome sometimes refers to a segment of a reference genome (eg, a chromosome or portion thereof, eg, one or more portions of a reference genome). Human genomes, human genome assemblies, and/or genomes from any other organism can be used as reference genomes. One or more human genomes, human genome assemblies, and genomes of other organisms are available at www. ncbi. nlm. nih. gov, National Center for Biotechnology Information. "Genome" refers to the complete genetic information of an organism or virus represented as nucleic acid sequences. As used herein, a reference sequence or reference genome is often an assembled or partially assembled genomic sequence obtained from one individual or multiple individuals. In some embodiments, the reference genome is an assembled or partially assembled genomic sequence obtained from one or more human individuals. In some embodiments, the reference genome comprises sequences assigned to chromosomes. The term "reference sequence" as used herein refers to one or more polynucleotide sequences of one or more reference samples. In some embodiments, the reference sequence comprises sequence reads obtained from a reference sample. In some embodiments, a reference sequence comprises sequence reads, assemblies of reads, consensus DNA sequences (e.g., sequence contigs), read densities, and/or read density profiles obtained from one or more reference samples. . A read density profile obtained from a reference sample is sometimes referred to herein as a reference profile. A read density profile obtained from a test sample and/or test subject is sometimes referred to herein as a test profile. In some embodiments, the reference sample is obtained from a reference subject that is substantially free of genetic mutations (eg, mutations in the gene of interest). In some embodiments, the reference sample is obtained from a reference subject that contains a known genetic mutation. The term "reference," as used herein, can refer to a reference genome, reference sequence, reference sample, and/or reference subject.

In certain embodiments, when the sample nucleic acid is derived from a pregnant female, the reference sequence is sometimes derived from neither the fetus, nor the fetal mother, nor the fetal father, referred to herein as " called an external reference. In some embodiments, a maternal reference can be prepared and used. When preparing a reference from a pregnant female (a "maternal reference sequence") based on an external reference, reads obtained from the pregnant female's DNA that are substantially free of fetal DNA are often Map against and collect external reference arrays. In certain embodiments, the external reference is derived from the DNA of an individual having substantially the same ethnicity as the pregnant female. A maternal reference sequence may not completely cover maternal genomic DNA (e.g., it may cover about 50%, 60%, 70%, 80%, 90% or more of maternal genomic DNA). some), the maternal reference may not be a perfect match to the maternal genomic DNA sequence (eg, the maternal reference sequence may contain multiple discrepancies).

In certain embodiments, mappability is assessed for genomic regions (eg, segments, genomic portions). Mappability is the mapping of a nucleotide sequence read to a portion of a reference genome, typically for which there are a specified number of mismatches, including, for example, 0, 1, 2 or more mismatches. The only thing is that it can be clearly aligned. In some embodiments, mappability is provided as a score or value generated by a suitable mapping algorithm or computer mapping software. For a given genomic region, the read-level mappability values obtained using a preset, read-length sliding window approach are averaged to obtain the expected mappability. can be estimated. Genomic regions containing stretches of unique nucleotide sequences sometimes have high mappability values.

Sequence reads can be mapped by a mapping module, or a machine that includes a mapping module, which generally maps the reads to a reference genome or segment thereof. The mapping module can map sequence reads by any suitable method known in the art. In some embodiments, a mapping module or machine that includes a mapping module is required to provide mapped sequence reads.

Counts Mapped sequence reads can be quantified to determine the number of reads mapped to a region or portion of the reference genome. In certain embodiments, reads that map to a reference genome, or a region, portion, or segment thereof, are referred to as counts. In some embodiments, the count number includes a value. In certain embodiments, the count value is determined by a mathematical process. The number of counts can be determined by any suitable method, manipulation, or mathematical treatment. In certain embodiments, counts are weighted, subtracted, filtered, normalized, adjusted, averaged, added or subtracted, or combinations thereof. In certain embodiments, counts are derived from sequence reads that are processed or manipulated by any suitable method, operation, or mathematical process described herein or known in the art. For example, counts are often normalized and/or weighted by one or more biases associated with sequence reads. In some embodiments, the counts are normalized and/or weighted according to the GC bias associated with the sequence reads. In some embodiments, counts are derived from raw reads of the sequence and/or filtered reads of the sequence. In some embodiments, one or more count numbers are not manipulated mathematically. The terms "raw count" and "raw counts" as used herein refer to one or more counts that have not been mathematically manipulated. .

In some embodiments, counts are determined for some or all of the sequence reads mapped to the reference genome, or a region, portion, or segment thereof. In certain embodiments, counts are determined from a pre-defined subset of mapped sequence reads. A pre-defined subset (eg, a selected subset) of mapped sequence reads can be defined or selected with the aid of any suitable characteristic or variable. In some embodiments, the predefined subset of mapped sequence reads can comprise 1 to n sequence reads, where n is generated from a test or reference sample. means a number equal to the sum of all sequence reads read.

Counts are often derived from sequence reads obtained from a subject (eg, a test subject). Counts are sometimes derived from sequence reads obtained from nucleic acid samples from pregnant females who give birth to a fetus. Nucleic acid sequence read counts are often counts representing both fetuses and fetal mothers (eg, of pregnant female subjects). In certain embodiments, when the subject is a pregnant female, some counts are from the fetal genome and some counts are from the maternal genome.

Read Density The read count (eg, weighted count) of an array is displayed as read density. Read densities are often determined and/or generated for one or more portions of a genome. In certain embodiments, read densities are determined and/or generated for one or more chromosomes. In some embodiments, read density comprises a quantitative measure of read counts of sequences mapped to portions of a reference genome. Read density can be determined by suitable procedures. In some embodiments, read density is determined by a suitable distribution and/or a suitable distribution function. Non-limiting examples of distribution functions include probability functions, probability distribution functions, probability density functions (PDF), kernel density functions (kernel density estimates), cumulative distribution functions, probability mass functions, discrete probability distributions, absolutely continuous univariate Any suitable distributions, such as distributions, or combinations thereof are included. In one particular embodiment, the PDF contains a kernel density function (kernel density estimate). Non-limiting examples of kernel density functions that can be used to generate local genomic bias estimates include uniform kernel density functions (uniform kernels), Gaussian kernel density functions (Gaussian kernels), triangular Kernel Density Function (Triangle Kernel), Biweight Kernel Density Function (Biweight Kernel), Tricube Kernel Density Function (Tricube Kernel), Triweight Kernel Density Function (Triweight Kernel), Cosine Kernel Function (Cosine Kernel), Epanechnikov Kernel density functions (Epanechnikov kernels), regular kernel density functions (regular kernels), etc., or combinations thereof. Read density is often a density estimate derived from a suitable probability density function. Density estimation is the construction of observational data-based estimates of the underlying probability density functions. In some embodiments, read density includes density estimation (eg, probability density estimation, kernel density estimation). Density estimation often includes kernel density estimation. In some embodiments, read density is a kernel density estimate determined according to a kernel density function. Read densities are often generated by a process that includes the step of generating a density estimate for each of one or more portions of the genome, each portion including a read count of the sequence. Read densities are often generated in terms of normalized and/or weighted counts mapped to parts. In some embodiments, each read mapped to a portion often contributes a value (eg, number of counts) equal to its weighting resulting from the normalization process described herein. In some embodiments, the read density for one or more portions is adjusted. Read density can be adjusted by any suitable method. For example, read densities for one or more portions can be weighted and/or normalized.

In some embodiments, the system includes distribution module 12 . The distribution module often generates and/or provides read densities (eg, 22, 24) for portions (eg, filtered portions) of the genome. The distribution module provides read densities, read density distributions 14, and/or associated measures of uncertainty (e.g., MAD, quantile) can be provided. The distribution module may accept, retrieve, and/or store sequence reads (eg, 1, 3, 5) and/or counts (eg, normalized counts of 11, weighted counts). The distribution module often accepts (eg, user input and user parameters for the portion), retrieves, generates, and/or stores the portion (eg, unfiltered or filtered portion). From time to time, the distribution module receives and/or retrieves portions (eg, filtered portions and/or selected portions 20) from filtering module 18. FIG. In some embodiments, the Distribution Module comprises code and/or source code (e.g., a collection of standard or custom scripts) that implements the functions of the Distribution Module, and/or one or more software packages (e.g., statistical It contains instructions (eg, algorithms, scripts) for the microprocessor in the form of a software package). In some embodiments, the distribution module includes code (eg, script) written in java, S, or R that leverages a suitable package (eg, S package, R package). A non-limiting example of a distribution module is provided in Example 2.

In some embodiments, a read density profile is determined. In some embodiments, a lead density profile includes at least one lead density, and often includes two or more lead densities (eg, a lead density profile often includes multiple lead densities). In some embodiments, the read density profile includes suitable quantitative values (eg, mean, median, z-score, etc.). A read density profile often includes one or more read density resulting values. A read density profile includes values resulting from one or more manipulations of read density based on one or more adjustments (eg, normalization). In some embodiments, the read density profile comprises unmanipulated read densities. In some embodiments, one or more read density profiles are derived from various aspects of a dataset comprising read densities or derivatives thereof (e.g., known in the art and/or described herein). the result of one or more mathematical and/or statistical data processing steps). In certain embodiments, the read density profile includes normalized read densities. In some embodiments, the read density profile includes adjusted read densities. In certain embodiments, the read density profiles are raw read densities (e.g., not manipulated, adjusted or normalized), normalized read densities, weighted read densities, filtered partial read density, read density z-score, read density p-value, read density integer value (e.g., area under the curve), read density mean, mean or median, principal component, etc., or combinations thereof. include. The read density of the read density profile and/or the read density profile are often associated with a measure of uncertainty (eg, MAD). In certain embodiments, the read density profile comprises a distribution of median read densities. In some embodiments, the read density profile includes multiple read density relationships (eg, fitted relationships, regressions, etc.). For example, sometimes a read density profile includes a relationship between read density (eg, read density value) and genomic location (eg, segment, location of segment). In some embodiments, the lead density profile is generated using static windowing, and in certain embodiments the lead density profile is generated using sliding windowing. The term “density read profile,” as used herein, is a mathematical and/or statistical manipulation of read density that can facilitate the identification of patterns and/or correlations in large sequence read data. refers to the results of

In some embodiments, the lead density profile is printed and/or displayed (eg, displayed as a visual representation, eg, plot or graph).

A read density profile often includes multiple data points, each data point representing a quantitative value of one or more read densities. Any suitable number of data points can be incorporated into the read density profile depending on the nature and/or complexity of the data set. In certain embodiments, the read density profile is 2 or more data points, 3 or more data points, 5 or more data points, 10 or more data points, 24 or more data points. points, 25 or more data points, 50 or more data points, 100 or more data points, 500 or more data points, 1000 or more data points, 5000 or more data points, It can include 10,000 or more data points, 100,000 or more data points, or 1,000,000 or more data points. In some embodiments, the data points are quantitative values and/or estimates of read counts of sequences mapped to or associated with one or more moieties. In some embodiments, the data points in the read density profile comprise the results of data manipulation of counts mapped to one or more portions. In certain embodiments, data points are often quantitative values and/or estimates of one or more read densities (eg, average read densities). A read density profile often includes multiple read densities associated with and/or mapped to multiple portions of a reference genome. In some embodiments, the read density profile comprises read densities from 2 to about 1,000,000 fractions. In some embodiments, 2 to about 500,000, 2 to about 100,000, 2 to about 50,000, 2 to about 40,000, 2 to about 30,000, 2 to about 20,000, 2 from about 10,000, 2 to about 5000, 2 to about 2500, 2 to about 1250, 2 to about 1000, 2 to about 500, 2 to about 250, 2 to about 100, or 2 to about 60 moieties The resulting read density determines the read density profile. In some embodiments, read densities from about 10 to about 50 fractions determine the read density profile.

In some embodiments, the read density profile corresponds to a set of portions (eg, a set of portions of a reference genome, a set of chromosomes, or a subset of portions of segments of chromosomes). In some embodiments, the read density profile includes read densities and/or counts associated with partial collections (eg, sets, subsets). In some embodiments, the lead density profile is determined for the lead densities of the continuous portion. In some embodiments, the contiguous portion includes gaps comprising segments of reference sequences and/or sequence reads not included in the density profile (eg, portions removed by filtering). Sometimes a contiguous portion (eg, a series of portions) represents contiguous segments of a genome or contiguous segments of a chromosome or gene. For example, two or more contiguous portions may represent a sequence assembly of DNA sequences longer than each portion when aligned by joining the portions end-to-end. For example, two or more contiguous portions can represent an intact genome, chromosome, gene, intron, exon, or segment thereof. Sometimes, a read density profile is determined from a collection (eg, set, subset) of contiguous and/or noncontiguous portions. In some cases, the read density profile is one or more parts weighted, excluded, filtered, normalized, adjusted, averaged, or derived as an average , may be added, subtracted, manipulated, or transformed by any combination thereof.

In some embodiments, the read density profile comprises read densities for portions of the genome containing genetic mutations. In some embodiments, the read density profile comprises read densities for portions of the genome that are free of genetic mutations (eg, portions of the genome that are substantially free of genetic mutations). In certain embodiments, the read density profile comprises read densities for portions of the genome containing genetic mutations and read densities for portions of the genome substantially free of genetic mutations.

Read density profiles are often determined for samples and/or references (eg, reference samples). Read density profiles are sometimes generated for entire genomes, for one or more chromosomes, or for portions or segments of genomes or chromosomes. In some embodiments, one or more read density profiles are determined for the genome or segments thereof. In some embodiments, the read density profile is a representation of the entire set of read densities of the sample, and in certain embodiments the read density profile is a representation of a portion or subset of the sample's read densities. . That is, the read density profile, in some cases, includes or is generated from read densities that represent data that has not been filtered to exclude any data, and the read density profile, in some cases, includes: Contains or is generated from data points that display data that has been filtered to exclude undesirable data.

In some embodiments, a read density profile is determined for a reference (eg, reference sample, training set). A read density profile for a reference is sometimes referred to herein as a reference profile. In some embodiments, a reference profile comprises read densities obtained from one or more references (eg, reference sequences, reference samples). In some embodiments, the reference profile comprises read densities determined for one or more (eg, a series) of known euploid samples. In some embodiments, the reference profile includes read densities of the filtered portion. In some embodiments, the reference profile includes read densities adjusted by one or more principal components.

In some embodiments, the system includes a profile generation module (eg, 26). Profile generation modules often receive, retrieve, and/or store read densities (eg, 22, 24). The profile generation module may generate read densities (e.g., adjusted, weighted, normalized, averaged, averaged, median, and/or integrated) from another suitable module (e.g., distribution module). read density) can be received and/or retrieved. The profile generation module can accept and/or retrieve read densities from a suitable source (eg, one or more reference subjects, a training set, one or more test subjects, etc.). The profile generation module often provides another suitable module (e.g., PCA statistics module 33, partial weighting module 42, scoring module 46) and/or to the user (e.g., by plotting, graphing and/or printing) generate and/or provide a read density profile (eg, 32, 30, 28) for the An example, or part thereof, of a profile generation module is provided in Example 2.

Portions In some embodiments, mapped sequence reads and/or counts are grouped together according to various parameters and referred to herein as “portions” or “a portion”. assigned to a particular segment and/or region of the reference genome called In some embodiments, the portion is an entire chromosome, a segment of a chromosome, a segment of a reference genome, a segment spanning multiple chromosomes, multiple segments of chromosomes, and/or combinations thereof. In some embodiments, portions are predefined based on certain parameters (eg, predetermined length, predetermined spacing, predetermined GC content, or any other suitable parameter). In some embodiments, portions are arbitrarily defined based on genome partitioning (e.g., partitioning by size, GC content, contiguous regions, contiguous regions of arbitrarily defined size, etc.). be. In some embodiments, portions are delineated based on one or more parameters, including, for example, sequence length or one or more specific characteristics. In some embodiments, the portion is based on a particular length of the genomic sequence. The portions may be approximately the same length, or the portions may be different lengths. In some embodiments, the portions are of approximately equal length. In some embodiments, portions of different lengths are adjusted or weighted. A portion can be of any suitable length. In some embodiments, portions are about 10 kilobases (kb) to about 100 kb, about 20 kb to about 80 kb, about 30 kb to about 70 kb, about 40 kb to about 60 kb, sometimes about 50 kb. In some embodiments, the portion is about 10 kb to about 20 kb. Portions are not limited to contiguous runs of sequences. Thus, portions may consist of contiguous and/or non-contiguous sequences.

In some embodiments, the portion comprises a window containing a preselected number of bases. The window may contain any suitable number of bases determined by the length of the portion. In some embodiments, the genome or segments thereof are partitioned into multiple windows. Windows that encompass regions of the genome may or may not overlap. In some embodiments, the windows are placed equidistant from each other. Some embodiments place the windows at different distances from each other. In certain embodiments, the genome or segments thereof are partitioned into multiple sliding windows, with the windows sliding progressively over the genome or segments thereof. The window can also be slid across the genome in any suitable increment or according to any numerical pattern or any non-thematic canonical sequence. In some embodiments, the window is about 100,000 bp or less, about 50,000 bp or less, about 25,000 bp or less, about 10,000 bp or less, about 5 ,000 bp or less, about 1,000 bp or less, about 500 bp or less, or about 100 bp or less increments. For example, the window may include approximately 100,000 bp and may be slid across the genome in 50,000 bp increments.

In some embodiments, the portion is a chromosome of interest, e.g., a particular chromosome such as a chromosome to assess genetic mutations (e.g., chromosomes 13, 18 and/or 21, or sex chromosome aneuploidy) can be a segment. A portion is not limited to a single chromosome. In some embodiments, the one or more portions comprise all or part of one chromosome, or all or part of two or more chromosomes. In some embodiments, one or more moieties may span across one, two or more chromosomes. Furthermore, portions can span contiguous or interspersed regions of multiple chromosomes. Portions can be genes, fragments of genes, regulatory sequences, introns, exons, and the like.

In some embodiments, certain regions of the genome are filtered prior to partitioning the genome or segments thereof into parts. Regions of the genome can be selected for exclusion from the partitioning process using any suitable method. Regions containing similar regions (eg, identical or homologous regions or sequences, eg, repeated regions) are often removed and/or filtered. Sometimes regions that cannot be mapped are excluded. In some embodiments, only unique regions are retained. The regions removed during partitioning may belong to a single chromosome or may span multiple chromosomes. In some embodiments, partitioned genomes are truncated and optimized for faster alignment, often allowing uniquely identifiable sequences to be focused. In some embodiments, the partitioning of the genome into regions (eg, regions beyond chromosomal limits) may be based on the gain of information generated in the context of classification. For example, measuring the significance of a particular genomic location for distinguishing between a group of subjects confirmed to be normal and a group of subjects confirmed to be abnormal (e.g., euploid and trisomy subjects, respectively). A p-value profile can be used to quantify information content. In some embodiments, partitioning of the genome into regions (e.g., regions beyond chromosomal limits) may be performed by any other criteria, e.g., speed/convenience in aligning reads, GC content ( high or low GC content), uniformity of GC content, other measures of sequence content (e.g. percentage of individual nucleotides, percentage of pyrimidines or purines, percentage of natural versus non-natural nucleic acids, percentage of methylated nucleotides, and CpG content), methylation status, melting temperature of the duplex, amenability to sequencing or PCR, measures of uncertainty assigned to individual parts of the reference genome, and/or It can be based on searches that target specific features, and the like.

A "segment" of a genome is sometimes a region comprising one or more chromosomes or portions of chromosomes. A "segment" is typically a portion of a genome that is distinct from a portion. A "segment" of a genome and/or chromosome sometimes is in a region of the genome or chromosome that is distinct from the part, sometimes shares no polynucleotides with the part, and sometimes contains polynucleotides within the part. A segment of a genome or chromosome often contains a greater number of nucleotides than a portion (e.g., a segment sometimes contains one or more portions), and a segment of a chromosome sometimes contains a smaller number of nucleotides than the portion. (eg, segments are sometimes inside parts).

Filtering Parts In certain embodiments, one or more parts (eg, genomic parts) are filtered out from consideration. In certain embodiments, one or more portions are filtered (eg, subjected to a filtering process), thereby presenting the filtered portions. In some embodiments, the filtering process excludes certain portions and retains portions (eg, subsets of portions). Portions that are retained after filtering are often referred to herein as filtered portions. In some embodiments, the reference genome portion is filtered. In some embodiments, portions of the reference genome filtered out are not included in determining the presence or absence of genetic variation (eg, chromosomal aneuploidy). In some embodiments, portions of chromosomes in the reference genome are filtered. In some embodiments, the portion associated with the read density (e.g., where the read density is the read density for the portion) is filtered out, and the read density associated with the excluded portion is the gene mutation (e.g., chromosomal aneuploidy) is not included in determining the presence or absence. In some embodiments, the read density profile comprises and/or consists of the read density of the filtered portion. Portions can be selected, filtered, and/or excluded from consideration using any suitable criteria and/or methods known in the art or described herein. Non-limiting examples of criteria used to filter portions include redundant data (e.g., redundant or duplicate mapped reads), non-informative data (e.g., reference genome portions with zero mapped counts), Reference genome portion with overrepresented or underrepresented sequence, GC content, noise data, mappability, counts, variability in counts, read density, variability in read density, uncertainty including measures of sexiness, measures of reproducibility, etc., or combinations of the foregoing. Portions are optionally filtered according to the distribution of counts and/or the distribution of read densities. In some embodiments, portions are filtered according to the distribution of counts and/or read densities, where the counts and/or read densities are obtained from one or more reference samples. One or more reference samples are sometimes referred to herein as a training set. In some embodiments, portions are filtered according to the distribution of counts and/or read densities, where the counts and/or read densities are obtained from one or more test samples. In some embodiments, the portion is filtered according to a measure of uncertainty about the read density distribution. In certain embodiments, the filtering process removes portions that support large deviations in read density. For example, a distribution of read densities (e.g., distribution of mean read density, mean read density, or median read density) if each read density in the distribution maps to the same portion; 5A distribution) can be determined. If each part of the genome is associated with a measure of uncertainty, the measure of uncertainty (eg, MAD) can be determined by comparing the distribution of read densities for multiple samples. Following the previous example, the portions can be filtered according to an uncertainty measure (eg, standard deviation (SD), MAD) associated with each portion and a predetermined threshold. FIG. 5B shows the distribution of MAD values for the part, determined according to the lead density distribution for multiple samples. A predetermined threshold is indicated by a vertical dashed line enclosing the range of acceptable MAD values. In the example of FIG. 5B, portions containing MAD values within the acceptable range are retained, and portions containing MAD values outside the acceptable range are filtered out of consideration. In some embodiments, in accordance with the preceding example, a filtering process removes portions containing read density values outside a predetermined uncertainty measure (e.g., median, mean, or mean read density) from consideration. often excluded. In some embodiments, portions containing read density values outside the interquartile range of the distribution (eg, median, mean, or average read density) are filtered out of consideration. In some embodiments, portions containing read density values that fall more than 2, 3, 4, or 5 times out of the interquartile range of the distribution are filtered out of consideration. In some embodiments, reads that deviate by more than 2 sigma, 3 sigma, 4 sigma, 5 sigma, 6 sigma, 7 sigma, or 8 sigma (eg, where sigma is in the range defined by the standard deviation) Portions containing density values are filtered out from consideration.

In some embodiments, the system includes filtering module 18 . A filtering module may filter the parts (e.g., parts of a given size and/or overlap, the position of the parts in the reference genome) and the read densities associated with the parts into another appropriate module (e.g., the distribution module 12). Read densities, which are often derived, are often received, retrieved, and/or stored. In some embodiments, the selected portions (eg, 20, eg filtered portions) are presented by a filtering module. In some embodiments, the filtering module is asked to present filtered portions and/or exclude portions from consideration. In certain embodiments, the filtering module excludes read densities from consideration if the read densities are associated with the excluded portion. The filtering module often presents the selected portion (eg filtered portion) to another suitable module (eg distribution module 21). A non-limiting example of a filtering module is presented in Example 3.

Bias Estimation Sequencing technology can be vulnerable to multiple sources of bias. In some cases, the sequencing bias is a local bias (eg, a local genomic bias). Local biases are often manifested at the level of sequence reads. The local genomic bias can be any suitable local bias. Non-limiting examples of local biases include sequence bias (e.g., GC bias, AT bias, etc.), DNase I sensitivity, entropy, repetitive sequence bias, chromatin structure bias, polymerase error rate bias. , batch sequence bias, inverted repeat bias, PCR-related bias, etc., or biases correlated with combinations thereof. In some embodiments, the source of local bias is undetermined or unknown.

In some embodiments, an estimate of local genomic bias is determined. Local genomic bias estimates are sometimes referred to herein as local genomic bias estimates. An estimate of local genomic bias can be determined for a reference genome, segment or portion thereof. In certain embodiments, a local genomic bias estimate is determined for one or more chromosomes in a reference genome. In some embodiments, an estimate of local genomic bias is determined for one or more sequence reads (eg, sequence reads for some or all of a sample). A local genomic bias estimate is determined for the sequence reads according to the local genomic bias estimate for the corresponding location and/or point in the reference (e.g., reference genome, chromosome in the reference genome). There are many things. In some embodiments, the local genomic bias estimate comprises a quantitative measure of sequence (eg, sequence reads, sequences of a reference genome) bias. Estimates of local genomic bias can be determined by suitable methods or mathematical treatments. In some embodiments, the estimate of local genomic bias is determined by a suitable distribution and/or a suitable distribution function (eg, PDF). In some embodiments, the estimate of local genomic bias comprises a quantitative representation of the PDF. In some embodiments, local genomic bias estimates (e.g., probability density estimates (PDEs), kernel density estimates) are combined with local bias content probability density functions (e.g., PDF: probability density function, eg kernel density function). In some embodiments, density estimation includes kernel density estimation. Estimates of local genomic bias are optionally expressed as the mean, mean, or median of the distribution. In some cases, estimates of local genomic bias are expressed as the sum or integral of an appropriate distribution (eg, area under a curve (AUC)).

A PDF (eg, kernel density function, eg, Epanechnikov kernel density function) often includes a bandwidth variable (eg, bandwidth). The bandwidth variable often defines the size and/or length of the window from which the probability density estimate (PDE) is derived when using PDF. The window from which the PDE is derived often contains a defined length of polynucleotide. In some embodiments, the window from which the PDE is derived is partial. Portions (eg, portion size, portion length) are often determined according to bandwidth variables. The length or size of the window used to determine the local genomic bias estimate by the bandwidth variable, from which the local genomic bias estimate is determined, the polynucleotide segment Determine the length or size of the window, which is the length of (eg, a contiguous segment of nucleotide bases). Non-limiting examples thereof include about 5 bases to about 100,000 bases, about 5 bases to about 50,000 bases, about 5 bases to about 25,000 bases, about 5 bases to about 10,000 bases, about 5 bases to about 5,000 bases, about 5 bases to about 2,500 bases, about 5 bases to about 1000 bases, about 5 bases to about 500 bases, about 5 bases to about 250 bases, about 20 bases to about 250 bases, etc. Any suitable bandwidth can be used to determine the PDE (eg, read density, estimate of local genomic bias (eg, GC density)), including the bandwidth of . In some embodiments, the local genomic bias estimate (e.g., GC density) is about 400 bases or less, about 350 bases or less, about 300 bases or less, about 250 bases or less. less than, about 225 bases or less, about 200 bases or less, about 175 bases or less, about 150 bases or less, about 125 bases or less, about 100 bases or less, about 75 bases or less , using a band width of about 50 bases or less, or about 25 bases or less. In certain embodiments, an estimate of local genomic bias (e.g., GC density) is the average read length of sequence reads obtained for a given subject and/or sample, average read length , the median lead length, or the bandwidth determined according to the maximum lead length. In some cases, an estimate of local genomic bias (e.g., GC density) is the mean read length, mean read length, median read length of the sequence reads obtained for a given subject and/or sample. length, or bandwidth approximately equal to the maximum lead length. In some embodiments, local genomic bias estimates (e.g., GC densities) are about , 110, 100, 90, 80, 70, 60, 50, 40, 30, 20, or about 10 bases.

Local genomic bias estimates can be determined with single-base resolution, but local genomic bias estimates (e.g., local GC content) can also be determined at low resolution. can. In some embodiments, a local genomic bias estimate is determined for local bias content. Estimates of local genomic bias (eg, determined using PDF) are often determined using windows. In some embodiments, the estimate of local genomic bias comprises the use of a window containing a preselected number of bases. In some cases, the window includes segments of consecutive bases. Optionally, the window includes a portion of one or more non-contiguous bases. In some cases, the window includes one or more portions (eg, genomic portions). The window size or length is often determined by the bandwidth and according to the PDF. In some embodiments, the window is about 10 times or more, 8 times or more, 7 times or more, 6 times or more, 5 times or more, 4 times or more the length of the bandwidth. more, three times or more, or about two times or more. If a PDF (eg, kernel density function) is used to determine the density estimate, the window is sometimes twice the length of the selected bandwidth. A window may contain any suitable number of bases. In some embodiments, the window is from about 5 bases to about 100,000 bases, from about 5 bases to about 50,000 bases, from about 5 bases to about 25,000 bases, from about 5 bases to about 10,000 bases, about 5 bases to about 5,000 bases, about 5 bases to about 2,500 bases, about 5 bases to about 1000 bases, about 5 bases to about 500 bases, about 5 bases to about 250 bases, or about 20 bases to about Contains 250 bases. In some embodiments, the genome or segments thereof are partitioned into multiple windows. Windows that encompass regions of the genome may or may not overlap. In some embodiments, the windows are placed equidistant from each other. Some embodiments place the windows at different distances from each other. In certain embodiments, the genome or segments thereof are partitioned into multiple sliding windows, with the windows sliding progressively over the genome or segments thereof. Each window of each increment contains an estimate of local genomic bias (eg, local GC density). The window can be slid across the genome in any suitable increment, can be slid according to any numerical pattern, or can be slid according to any unthematic canonical sequence. In some embodiments, about 10,000 bp or more, about 5,000 bp or more, about 2,500 bp or more are used to determine local genomic bias estimates. greater than, about 1,000 bp or more, about 750 bp or more, about 500 bp or more, about 400 bases or more, about 250 bp or more, about 100 bp or more, about 50 bp or more, or about 25 bp Or slide the window in increments beyond that. In some embodiments, about 25, 24, 23, 22, 21, 20, 19, 18, 17, 16, 15, across the genome or segments thereof, to determine an estimate of local genomic bias. , 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, or about 1 bp increments. For example, to determine an estimate of local genomic bias, the window can include approximately 400 bp (eg, a bandwidth of 200 bp) and can be slid across the genome in 1 bp increments. In some embodiments, a kernel density function and a bandwidth of approximately 200 bp are used to determine local genomic bias estimates for each base within the genome or segment thereof.

In some embodiments, the estimate of local genomic bias is local GC content and/or an indication of local GC content. The term "local" as used herein (e.g., local bias, local bias estimate, local bias content, local genomic bias, local GC content, etc. ) refers to a polynucleotide segment of 10,000 bp or less. In some embodiments, the term "local" refers to a or less, 175 bp or less, 150 bp or less, 100 bp or less, 75 bp or less, or 50 bp or less. Local GC content is a representation (e.g., mathematical representation, quantitative representation) of GC content for a local segment of a genome, sequence read, sequence read assembly (e.g., contig, profile, etc.) There are many things. For example, the local GC content may be an estimate of the local GC bias, or it may be the local GC density.

One or more GC densities are often determined for a reference or sample (eg, test sample) polynucleotide. In some embodiments, GC density is a representation (eg, a mathematical representation, a quantitative representation) of local GC content (eg, for polynucleotide segments of 5000 bp or less). In some embodiments, GC density is an estimate of local genomic bias. GC density can be determined using suitable processes described herein and/or known in the art. GC density can be determined using a suitable PDF (eg, kernel density function (eg, Epanechnikov kernel density function, eg, see FIG. 1)). In some embodiments, the GC density is a PDE (eg kernel density estimate). In certain embodiments, GC density is defined by the presence or absence of one or more guanine (G) nucleotides and/or cytosine (C) nucleotides. Conversely, in some embodiments, GC density can also be defined by the presence or absence of one or more adenine (A) and/or thymidine (T) nucleotides. In some embodiments, GC density for local GC content is determined by whole genomes or segments thereof (e.g., autosomes, sets of chromosomes, single chromosomes, genes; see, e.g., Figure 2). Normalize according to the GC density determined for One or more GC densities can be determined for polynucleotides in a sample (eg, test sample) or reference sample. GC density is often determined for a reference genome. In some embodiments, GC density is determined for sequence reads according to the reference genome. The GC density of a read is often determined according to the GC density determined for the corresponding position and/or point in the reference genome to which the read is mapped. In some embodiments, GC densities determined for locations on the reference genome are assigned and/or presented for reads, where the reads or segments thereof map to locations on the same reference genome. be. Any suitable method can be used to determine the location of mapped reads on the reference genome for the purpose of generating GC densities for the reads. In some embodiments, the median point of mapped reads determines the position on the reference genome from which to determine GC density for reads derived therefrom. For example, if the median point of a read maps to chromosome 12 at base number x of the reference genome, then the GC density of the read is the point located on chromosome 12 at or near base number x of the reference genome. It is often presented as the GC density determined by the kernel density estimate for . In some embodiments, GC density is determined for some or all base points of the read according to the reference genome. Optionally, the GC density of a read comprises the average, sum, median, or integral of two or more GC densities determined for multiple base points on the reference genome.

In some embodiments, estimates of local genomic bias (eg, GC density) are quantified and/or presented as values. Estimates of local genomic bias (eg, GC density) are sometimes expressed as mean, average, and/or median. An estimate of local genomic bias (eg, GC density) is sometimes expressed as the maximum peak height of the PDE. In some cases, estimates of local genomic bias (eg, GC density) are expressed as sums or integrals of appropriate PDEs (eg, area under the curve (AUC)). In some embodiments, the GC density includes kernel weights. In certain embodiments, the GC density of a read comprises a value approximately equal to the mean, mean, sum, median, maximum peak height, or integral of kernel weights.

Biased Frequencies Biased frequencies are optionally determined according to one or more local genomic bias estimates (eg, GC density). A bias frequency is optionally a count or sum of the number of occurrences of local genomic bias estimates for a sample, a reference (e.g., a reference genome, a reference sequence, a chromosome in a reference genome), or a portion thereof. is. A biased frequency is optionally a count or sum of the number of occurrences of a local genomic bias estimate (e.g., each local genomic bias estimate) for a sample, reference, or portion thereof. is. In some embodiments, the biased frequency is the GC density frequency. GC density frequencies are often determined according to one or more GC densities. For example, GC density frequency may indicate the number of times a GC density of value x is displayed over the entire genome or a segment thereof. The bias frequency is often the distribution of local genomic bias estimates, where the number of occurrences of each local genomic bias estimate is denoted as the bias frequency (e.g., FIG. 3 (see ). Biased frequencies are optionally mathematically manipulated and/or normalized. Biased frequencies can be mathematically manipulated and/or normalized by any suitable method. In some embodiments, the biased frequency is an estimate of each local genomic bias (e.g., an autosome, a subset of chromosomes, a single chromosome, or normalize according to the indication (e.g., fraction, percentage) of the reads). A biased frequency can be determined for an estimate of the local genomic bias of some or all of the samples or references. In some embodiments, biased frequencies can be determined for local genomic bias estimates for some or all sequence reads of a test sample.

In some embodiments, the system includes biased density module 6 . The Bias Density Module accepts, retrieves, and/or stores the mapped sequence reads 5 and reference sequences 2 in any suitable format and provides local genomic bias estimates, local genomic bias distributions , biased frequencies, GC densities, GC density distributions, and/or GC density frequencies (together represented by box 7) can be generated. In some embodiments, the biased density module transfers data and/or information (eg, 7) to another appropriate module (eg, relationship module 8).

Relationships In some embodiments, one or more relationships are generated between local genomic bias estimates and bias frequencies. As used herein, the term "relationship" refers to a mathematical and/or graphical relationship between two or more variables or values. Relationships can be generated by suitable mathematical and/or graphical processes. Non-limiting examples of relationships include mathematical and/or graphical representations of functions, correlations, distributions, linear or non-linear equations, straight lines, regressions, fitted regressions, etc., or combinations thereof. In some cases, a relationship includes a matched relationship. In some embodiments, the fitted relationship comprises a fitted regression. In some cases, a relationship is two or more variables or values and includes weighted variables or weighted values. In some embodiments, the relationship comprises a fitted regression, where one or more variables or values of the relationship are weighted. In some cases the regression is fitted in a weighted fashion. In some cases, the regression fits unweighted. In certain embodiments, generating relationships includes plotting or graphing.

In some embodiments, a suitable relationship is determined between the local genomic bias estimate and the bias frequency. In some embodiments, the sample bias relationship is presented by generating a relationship between (i) the local genomic bias estimate and (ii) the bias frequency for the sample. In some embodiments, the reference bias relationship is presented by generating a relationship between (i) the local genomic bias estimate and (ii) the bias frequency for the reference. In one particular embodiment, a relationship is created between GC density and GC density frequency. In some embodiments, the sample GC density relationship is presented by generating a relationship between (i) GC density and (ii) GC density frequency for the sample. In some embodiments, the reference GC density relationship is presented by generating a relationship between (i) GC density and (ii) GC density frequency for the reference. In some embodiments, when the local genomic bias estimate is GC density, the sample bias relation is the sample GC density relation and the reference bias relation is the reference GC density relation. The GC density of the reference GC density relationship and/or sample GC density relationship is often an indication (eg, a mathematical or quantitative indication) for local GC content. In some embodiments, the relationship between the local genomic bias estimate and the bias frequency comprises a distribution. In some embodiments, the relationship between the local genomic bias estimate and the bias frequency comprises a fitted relationship (eg, a fitted regression). In some embodiments, the relationship between the local genomic bias estimate and the bias frequency comprises a linear fit regression or a non-linear fit regression (eg, polynomial regression). In certain embodiments, the relationship between the local genomic bias estimate and the bias frequency comprises a weighted relationship, wherein the local genomic bias estimate and/or the bias frequency are appropriately weighted by In some embodiments, weighted fit relationships (eg, weighted fits) can be obtained by processes including quartile regression, parameterized probability distributions, or empirical distributions with interpolation. In certain embodiments, the relationship between the local genomic bias estimate and the bias frequency for the test sample, reference, or portion thereof comprises polynomial regression, wherein the local genomic bias estimate is Values are weighted. In some embodiments, the weighted fitting model includes weighting the distribution values. The distribution values can be weighted by suitable processing. In some embodiments, values near the tail of the distribution are given less weight than values near the median of the distribution. For example, for a distribution of local genomic bias estimates (e.g., GC densities) and bias frequencies (e.g., GC density frequencies), the weight for a given local genomic bias estimate is where local genomic bias estimates containing bias frequencies close to the mean of the distribution are larger than local genomic bias estimates containing bias frequencies far from the mean. give weight.

In some embodiments, the system includes relationship module 8 . A relationship module allows the generation of relationships as well as the functions, coefficients, constants, and variables that define the relationships. A relationship module may receive, store and/or retrieve data and/or information (eg 7) from an appropriate module (eg bias density module 6) to generate a relationship. Relation modules often generate and compare distributions of local genomic bias estimates. A relationship module allows the data sets to be compared and optionally to generate regression and/or fitted relationships. In some embodiments, the relational module compares one or more distributions (e.g., distributions of local genomic bias estimates of the sample and/or reference) for sequence read counts. The weighting factors and/or weight assignments 9 are submitted to another suitable module (eg a bias correction module). Optionally, the relationship module presents the normalized sequence read counts directly to the distribution module 21, where the counts are normalized according to relationships and/or comparisons.

Generating Comparisons and Their Use In some embodiments, processing to reduce local bias in sequence reads includes normalizing sequence read counts. Sequence read counts are often normalized according to comparison of the test sample to the reference. For example, in some cases, the sequence read count is an estimate of the local genomic bias of the sequence reads of the test sample versus the local genomic bias of the reference (e.g., the reference genome or a portion thereof). Normalize by comparing with the estimated value. In some embodiments, the sequence read count is obtained by comparing the bias frequency of the test sample local genomic bias estimate to the bias frequency of the reference local genomic bias estimate. Normalize by In some embodiments, sequence read counts are normalized by comparing the sample-biased relation to the reference-biased relation, thereby generating a comparison.

Sequence read counts are often normalized according to a comparison of two or more relationships. In certain embodiments, two or more relationships are compared, thereby reducing local bias (e.g., normalizing counts) in sequence reads. Present. Any suitable method can compare two or more relationships. In some embodiments, the comparison includes adding the second relationship to the first relationship, subtracting the second relationship from the first relationship, multiplying the first relationship by the second relationship and/or dividing the first relationship by the second relationship. In certain embodiments, comparing two or more relationships includes using suitable linear and/or non-linear regression. In certain embodiments, comparing two or more relationships includes a suitable polynomial regression (eg, third order polynomial regression). In some embodiments, the comparison includes adding the second regression to the first regression, subtracting the second regression from the first regression, multiplying the first regression by the second regression and/or dividing the first regression by the second regression. In some embodiments, two or more relationships are compared by a process that includes a multiple regression inference framework. In some embodiments, two or more relationships are compared by processing including suitable multivariate analysis. In some embodiments, two or more relationships are compared by processes including basis functions (e.g., blending functions, e.g., polynomial basis, Fourier basis, etc.), splines, radial basis functions, and/or wavelets. .

In certain embodiments, distributions of local genomic bias estimates, including biased frequencies for test samples and references, are compared by a process involving polynomial regression, where local genomic bias The estimates of are weighted. In some embodiments, polynomial regression is performed by (i) each of the ratios comprising the bias frequency of the reference local genomic bias estimate and the bias frequency of the sample local genomic bias estimate and (ii) an estimate of local genomic bias. In some embodiments, polynomial regression is performed by (i) the ratio of the bias frequency of the reference local genomic bias estimate to the bias frequency of the sample local genomic bias estimate; ) to generate local genomic bias estimates. In some embodiments, the comparison of the distributions of the local genomic bias estimates for the test sample and the reference reads is the logarithm of the bias frequency of the local genomic bias estimates for the reference and the sample. Including determining a ratio (eg, log ₂ ratio). In some embodiments, the comparison of the distributions of the local genomic bias estimates is the log ratio (e.g., log ₂ ratio) of the bias frequency of the local genomic bias estimates relative to the reference: Including dividing by the log ratio (eg, log ₂ ratio) of the bias frequency of the local genomic bias estimate for the sample (see, eg, Example 1 and FIG. 4).

Normalizing the counts according to the comparison typically adjusts some counts but not others. Normalizing the counts adjusts the total counts in some cases and does not adjust the read counts in any sequence in some cases. In some cases, the counts for sequence reads are normalized by a process that involves determining weighting factors, and in other cases, the process does not involve direct generation and exploitation of weighting factors. Normalizing the counts according to the comparison optionally includes determining a weighting factor for the read counts of each sequence. The weighting factors are sequence read specific and are often applied to the read counts for a specific sequence. Weighting factors are often determined according to a comparison of two or more bias relationships (eg, a sample bias relationship compared to a reference bias relationship). The normalized count number is often determined by adjusting the count number according to a weighting factor. Adjusting the number of counts according to the weighting factor may optionally include adding the weighting factor to the number of counts for array reads, subtracting the weighting factor from the number of counts for array reads, or counts for array reads. Multiplying the number by a weighting factor and/or dividing the number of counts for array reads by the weighting factor. Weighting factors and/or normalized counts are optionally determined from regression (eg, a regression line). The normalized counts are optionally the bias frequency of the reference local genomic bias estimate (e.g., the reference genome, the chromosomes in the reference genome) and the test sample local genomic bias estimate. Directly from the regression line (eg, the fitted regression line) resulting from the comparison of the values with the frequency of bias. In some embodiments, each count of reads in a sample is expressed as (i) the bias frequency of the local genomic bias estimate of the read, (ii) the local genomic bias estimate of the reference. Presented as normalized count values according to the comparison compared to the biased frequency. In certain embodiments, the sequence read counts obtained for a sample are normalized to reduce bias in sequence reads.

Optionally, the system includes a bias correction module 10. FIG. In some embodiments, the function of bias correction module is performed by relationship modeling module 8 . The bias correction module can accept, retrieve, and/or store mapped sequence reads and weighting factors (eg, 9) from appropriate modules (eg, relation module 8, compression module 4). In some embodiments, the bias correction module presents counts to mapped reads. In some embodiments, the bias correction module applies weight assignments and/or bias correction factors to the sequence read counts, thereby presenting normalized and/or adjusted counts. The bias correction module often presents the normalized counts to another suitable module (eg distribution module 21).

In certain embodiments, normalizing the counts comprises factorizing one or more features in addition to the GC density and normalizing the sequence read counts. In certain embodiments, normalizing the counts comprises factorizing one or more different local genomic bias estimates and normalizing the sequence read counts. including. In certain embodiments, sequence read counts are weighted according to weightings determined according to one or more characteristics (eg, one or more biases). In some embodiments, counts are normalized according to one or more combined weights. In some cases, factorizing one or more features and/or normalizing counts according to one or more combined weights is via processing that includes the use of multivariate models. Any suitable multivariate model can be used to normalize the counts. Non-limiting examples of multivariate models include multivariate linear regression, multivariate quartile regression, multivariate interpolation of empirical data, nonlinear multivariate models, etc., or combinations thereof.

In some embodiments, the system includes multivariate correction module 13 . The multivariate correction module may perform the functions of the bias density module 6, the relationship module 8, and/or the bias correction module 10 multiple times, thereby adjusting the counts for multiple biases. In some embodiments, the multivariate correction module includes one or more of bias density module 6, relationship module 8, and/or bias correction module 10. Optionally, the multivariate correction module presents the normalized counts 11 to another appropriate module (eg distribution module 21).

Weighted Portion In some embodiments, the portion is weighted. In some embodiments, one or more portions are weighted, thereby presenting weighted portions. Weighted parts optionally remove partial dependencies. Portions can be weighted by appropriate processing. In some embodiments, one or more portions are weighted by an eigen function (or eigenfunction). In some embodiments, the eigenfunction includes replacing the portion with an orthogonal eigenportion. In some embodiments, the system includes partial weighting module 42 . In some embodiments, a weighting module receives, retrieves, and/or stores read densities, read density profiles, and/or adjusted read density profiles. In some embodiments, weighted portions are presented by a partial weighting module. In some embodiments, a weighting module is called upon to weight the portions. The weighting module can weight the portions according to one or more weighting methods known in the art or described herein. The weighting module often presents the weighted portion to another suitable module (eg, scoring module 46, PCA statistics module 33, profile generation module 26, etc.).

Principal Component Analysis In some embodiments, the read density profile (eg, the read density profile of the test sample (eg, FIG. 7A)) is adjusted according to principal component analysis (PCA). One or more reference sample read density profiles and/or test subject read density profiles can be adjusted according to the PCA. Read density profiles for genomes, portions of genomes, chromosomes, or segments of chromosomes can be prepared by PCA. Herein, removal of bias from a read density profile, via PCA-related processing, is sometimes referred to as adjusting the profile. PCA can be performed by any suitable PCA method or variations thereof. Non-limiting examples of PCA methods are Canonical Correlation Analysis (CCA), Karhunen-Loeve Transform (KLT), Hotelling Transform, EigenOrthogonal Decomposition (POD), Singular Value Decomposition (SVD) of X, Eigenvalue of XTX Decomposition (EVD), factor analysis, Eckert Young's theorem, Schmidt-Mirski's theorem, empirical orthogonal function (EOF), empirical eigenfunction decomposition, empirical component analysis, quasi-harmonic modes, spectral decomposition, empirical modal analysis, etc. , including variations or combinations thereof. PCA often identifies one or more biases in the read density profile. Herein, biases identified by PCA are sometimes referred to as principal components. In some embodiments, one or more biases can be eliminated by adjusting the read density profile according to one or more principal components using suitable methods. The read density profile is obtained by adding one or more principal components to the read density profile, subtracting one or more principal components from the read density profile, and multiplying the read density profile by one or more principal components. and/or by dividing the read density profile by one or more principal components. In some embodiments, one or more biases can be removed from the read density profile by subtracting one or more principal components from the read density profile. Bias in read density profiles are often identified and/or quantified by PCA of the profile, whereas principal components are often subtracted from the profile at the level of read density. Read density profile biases or features identified and/or quantified by profile PCA include, but are not limited to, fetal sex, sequence bias (e.g., guanine and cytosine (GC) bias), fetal fraction , bias correlated with DNase I sensitivity, entropy, repetitive bias, chromatin structure bias, polymerase error rate bias, batch bias, inverted repeat bias, PCR amplification bias, and hidden copy number. Mutations included.

PCA often identifies one or more principal components. In some embodiments, PCA extracts the 1st, 2nd, 3rd, 4th, 5th, 6th, 7th, 8th, 9th, and 10th or higher order principal components identify. In certain embodiments, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10 or more principal components are used to tune the profile. In one particular embodiment, five principal components are used to tune the profile. Principal components are often used to adjust profiles in order of their appearance in PCA. For example, if three principal components are subtracted from the read density profile, the first, second, and third principal components are used. In some cases, the biases identified by the principal components include features of the profile that are not used to adjust the profile. For example, PCA can identify genetic mutations (eg, aneuploidies, deletions, rearrangements, insertions) and/or sex differences (eg, seen in FIG. 6C) as major components. Therefore, in some embodiments, one or more principal components are not used to adjust the profile. For example, in some cases, the first, second, and fourth principal components are used to adjust the profile, where the third principal component is not used to adjust the profile. Principal components can be obtained from PCA using any suitable sample or reference. In some embodiments, the principal component is obtained from a test sample (eg, test subject). In some embodiments, principal components are obtained from one or more references (eg, reference samples, reference sequences, reference sets). For example, as shown in Figure 6, PCA was obtained from a training set (Figure 6A) containing multiple samples resulting in the identification of a first principal component (Figure 6B) and a second principal component (Figure 6C). It is performed on the median read density profile In some embodiments, principal components are obtained from a set of subjects known to lack mutations in the gene of interest. In some embodiments, principal components are obtained from a known euploid set. Principal components are often identified according to PCA performed using one or more read density profiles (eg, a training set) of reference. One or more principal components obtained from the reference are often subtracted from the test subject's read density profile (eg, FIG. 7B), thereby presenting an adjusted profile (eg, FIG. 7C).

次いで、主成分は以下のように計算される：

最終的に、各試料のＰＣＡ調整プロファイルは、以下のように計算され得る：

In some embodiments, the system includes a PCA statistics module 33. The PCA statistics module can accept and/or retrieve read density profiles from another suitable module (eg, profile generation module 26). PCA is often performed by a PCA statistics module. A PCA statistics module accepts, retrieves, and/or stores read density profiles and often processes read density profiles from a reference set 32, a training set 30, and/or one or more test subjects 28. The PCA statistics module can generate and/or present principal components and/or adjust the read density profile according to one or more principal components. Adjusted read density profiles (eg, 40, 38) are often provided by the PCA statistics module. The PCA statistics module can present and/or forward the adjusted read density profile (eg, 38, 40) to another appropriate module (eg, partial weighting module 42, scoring module 46). In some embodiments, gender determination 36 can be presented by the PCA statistics module. Gender determination is the determination of the sex of a fetus, optionally determined according to PCA and/or according to one or more principal components. In some embodiments, the PCA statistics module includes some, all, or one modification of the R code shown below. R code for computing principal components generally begins with cleaning the data (eg, subtracting medians, filtering parts, and trimming extrema).

The principal components are then calculated as follows:

Ultimately, the PCA adjusted profile for each sample can be calculated as follows:

Comparing Profiles In some embodiments, determining an outcome comprises a comparison. In certain embodiments, read density profiles, or portions thereof, are leveraged to present outcomes. In certain embodiments, read density profiles for genomes, portions of genomes, chromosomes, or segments of chromosomes are leveraged to provide outcomes. In some embodiments, determining an outcome (eg, determining the presence or absence of a genetic mutation) comprises comparing two or more read density profiles. Comparing read density profiles often involves comparing read density profiles made for selected segments of the genome. For example, a test profile is often compared to a reference profile, and the test profile and the reference profile were determined for segments of the genome that are substantially the same segment (eg, the reference genome). Comparing read density profiles optionally includes comparing two or more subsets of portions of the read density profiles. A subset of portions of the read density profile can represent segments of the genome (eg, chromosomes or segments thereof). A read density profile can include any amount of subsets of portions. Optionally, the read density profile includes 2 or more, 3 or more, 4 or more, or 5 or more subsets. In certain embodiments, the read density profile includes two subsets of portions, where each portion represents adjacent segments of the reference genome. In some embodiments, the test profile can be compared to a reference profile, wherein both the test profile and the reference profile comprise a first subset of portions and a second subset of portions, wherein , the first subset and the second subset represent different segments of the genome. A subset of portions of the read density profile can contain genetic mutations, and other subsets of the portions are optionally substantially free of genetic mutations. Optionally, all subsets of the portion of the profile (eg, test profile) are substantially free of genetic mutations. Optionally, all subsets of the portion of the profile (eg, test profile) contain mutations of the gene. In some embodiments, a test profile may include a first subset of portions containing genetic mutations and a second subset of portions substantially free of genetic mutations.

In some embodiments, the methods described herein include preforming a comparison (eg, comparing a test profile to a reference profile). Suitable methods can compare two or more data sets, two or more relationships, and/or two or more profiles. Non-limiting examples of statistical methods suitable for comparing datasets, relationships, and/or profiles include the Behrens-Fischer method, the Bootstrap method, the Fisher method for combining independent significance tests, the Neiman-Pearson test. , confirmatory data analysis, exploratory data analysis, exact test, F-test, Z-test, T-test, calculation and/or comparison of uncertainty measures, null hypothesis, counternull, etc., chi-square test, Including omnibus tests, calculating and/or comparing levels of significance (eg, statistical significance), meta-analyses, multivariate analyses, regression, simple linear regression, robust linear regression, etc., or combinations of the foregoing. In certain embodiments, comparing two or more data sets, relationships, and/or profiles includes determining and/or comparing measures of uncertainty. As used herein, "uncertainty measure" refers to a measure of significance (e.g., statistical significance), a measure of error, a measure of variance, a measure of reliability, etc., or a combination thereof. Point. The measure of uncertainty can be a value (eg, threshold) or a range of values (eg, interval, confidence interval, Bayesian confidence interval, threshold range). Non-limiting examples of uncertainty measures include p-values, suitable measures of deviation (e.g., standard deviation, sigma, absolute deviation, mean absolute deviation, etc.), suitable measures of error (e.g., standard error, squared mean error, root mean square error, etc.), appropriate measures of variance, appropriate standard scores (e.g. standard deviation, cumulative percentage, percentile equivalents, Z-score, T-score, R-score, standard 9-step method). (stanin), percent stanine, etc.), etc., or combinations thereof. In some embodiments, determining the level of significance includes determining a measure of uncertainty (eg, p-value). In certain embodiments, two or more data sets, relationships, and/or profiles are analyzed by multiple (e.g., two or more) statistical methods (e.g., least squares regression, principal component analysis). , linear discriminant analysis, quadratic discriminant analysis, bagging, neural networks, support vector machine models, random forests, classification tree models, K nearest neighbors, logistic regression and/or LOESS smoothing), and/or any suitable mathematical manipulation. and/or can be analyzed and/or compared by utilizing statistical manipulations (eg, referred to herein as manipulations).

In certain embodiments, comparing two or more read density profiles includes determining and/or comparing uncertainty measures for the two or more read density profiles. In some cases, read density profiles and/or associated uncertainty measures are compared to facilitate interpretation of mathematical and/or statistical manipulations of data sets and/or to present outcomes. Optionally, the read density profile generated for the test subject is compared to read density profiles generated for one or more references (eg, reference samples, reference subjects, etc.). In some embodiments, the outcome is presented by comparing a read density profile derived from the test subject to a read density profile derived from a reference for a chromosome, portion, or segment thereof, where the read of the reference Density profiles are obtained from a set of reference subjects (eg, references) that are known not to carry mutations in the gene. In some embodiments, the outcome is presented by comparing a read density profile derived from the test subject to a read density profile derived from a reference for a chromosome, portion, or segment thereof, where the read of the reference Density profiles are obtained from a set of reference subjects known to carry specific gene mutations (eg, chromosomal aneuploidy, trisomy).

In certain embodiments, the read density profile of the tested subject is compared to a predetermined value indicative of the absence of mutations in the gene, optionally with one or more Deviations from predetermined values in genomic locations (eg, portions). For example, in a test subject (e.g., a subject at risk for or suffering from a medical condition associated with a mutation in a gene), the read density profile is is expected to be significantly different from the read density profile of the reference (eg, reference sequence, reference subject, reference set). The read density profile of the test subject is substantially the same as the read density profile of the reference (e.g., reference sequence, reference subject, reference set) for selected portions where the test subject does not contain the mutation of the gene in question There are many things. Read density profiles are often compared to predetermined thresholds and/or threshold ranges (see, eg, FIG. 8). The term "threshold" as used herein is calculated using a qualitative data set and is the diagnostic limit for genetic mutations (e.g., copy number variations, aneuploidies, chromosomal abnormalities, etc.). Any number that can be used as In certain embodiments, the threshold is exceeded by the results obtained by the methods described herein and the subject is diagnosed with a genetic mutation (eg, trisomy). In some embodiments, the threshold value or threshold value range is determined through mathematical and/or statistical manipulation of sequence read data (e.g., from a reference and/or subject). is often calculated as A predetermined threshold or range of thresholds indicative of the presence or absence of a genetic mutation can vary while still providing a useful outcome for determining the presence or absence of a genetic mutation. In certain embodiments, a read density profile including normalized read densities and/or normalized counts is generated to facilitate classification and/or presentation of outcomes. Outcomes can be presented based on plots of read density profiles that include normalized counts (eg, using such plots of read density profiles).

In some embodiments, the system includes scoring module 46 . The scoring module converts the read density profile (e.g., adjusted normalized read density profile) from another suitable module (e.g., profile generation module 26, PCA statistics module 33, partial weighting module 42, etc.). It may be received, retrieved, and/or stored. The scoring module can accept, retrieve, store, and/or compare two or more read density profiles (eg, test profile, reference profile, training set, test subject). The scoring module allows scores (e.g. plots, profile statistics, comparisons (e.g. differences between two or more profiles), Z-scores, uncertainty measures, decision zones, sample decisions 50 (e.g. gene (determining the presence or absence of mutations in phenotypes), and/or outcomes). A scoring module can present the score to the end user and/or another suitable module (eg, display, printer, etc.). In some embodiments, the scoring module is R code, shown below, for calculating a chi-square statistic for a specific test (e.g., high chromosome 21 count). Contains modifications of part, all, or one of R code containing functions.
The three parameters are
x = sample read data (part x sample)
m = median for parts y = test vector (e.g. false for all parts except for chromosome 21 which is true)
is.

Experimental Conditions In certain embodiments, the principal component normalization process can be adjusted for biases associated with experimental conditions. Data processing considering experimental conditions is described, for example, in International Patent Application Publication No. WO2013/109981, the entire contents of which are incorporated herein by reference, including all documents, tables, formulas, and figures. It is

In certain cases, samples may be affected by common experimental conditions. Samples processed at substantially the same time or using substantially the same conditions and/or reagents sometimes were processed at different times and/or at the same time using different conditions and/or reagents Exhibits data variability (eg, bias) induced by similar experimental conditions (eg, common experimental conditions) when compared to other samples. There are often practical considerations that limit the number of samples that can be prepared, processed, and/or analyzed at any given time during an experimental procedure. In certain embodiments, the timeframe for processing samples from raw materials to generate outcomes is sometimes days, weeks, or even months. Due to the time between isolation and final analysis, high-throughput experiments that analyze large numbers of samples generate batch effects or experimental condition-induced data variability. Experimental condition-induced data variability often includes any data variability that is the result of sample isolation, storage, preparation, and/or analysis. Non-limiting examples of experimental condition-induced variability include sequence overrepresentation or underrepresentation; noisy data; spurious or outlier data points, reagent effects, personnel effects, laboratory condition effects, etc. Flow cell-based variability and/or plate-based variability are included. Experimental condition-induced variability sometimes occurs in subpopulations of samples in a dataset (eg, batch effect). Batches often include samples processed using substantially the same reagents, samples processed in the same sample preparation plate (e.g., sample preparation; e.g., microwell plates used for nucleic acid isolation), the same Samples developed for analysis in a development plate (e.g., a microwell plate used to organize samples prior to loading onto a flow cell), samples processed at substantially the same time, processed by the same personnel and/or processed under substantially the same experimental conditions (eg, temperature, _CO2 level, ozone level, etc., or a combination thereof). Experimental condition batch effects were sometimes analyzed in the same flow cell, prepared in the same reagent plate or microwell plate, and/or developed for analysis in the same reagent plate or microwell plate (e.g., for sequencing). (preparing a nucleic acid library) to influence the sample. Additional sources of variability can include the quality of the nucleic acid isolated, the amount of nucleic acid isolated, time to storage after nucleic acid isolation, time during storage, storage temperature, etc., and combinations thereof. . Variability in data points within a batch (e.g., subpopulations of samples in a dataset that are processed at the same time and/or using the same reagents and/or experimental conditions) is sometimes seen between batches. greater than the variability of the data points. This data variability sometimes includes spurious or outlier data whose magnitude can lead to the interpretation of some or all other data in the data set. Part or all of the data set may include data processing steps described herein and known in the art; e.g. Normalization for deviation can be used to adjust for experimental conditions. Data processing considering experimental conditions is described, for example, in International Patent Application Publication No. WO2013/109981, the entire contents of which are incorporated herein by reference, including all documents, tables, formulas, and figures. It is

Aneuploidy Detection Using Comparisons In some embodiments, principal component normalization processing is used in conjunction with methods for determining the presence or absence of aneuploidy according to comparisons. Aneuploidy detection using comparison is described, for example, in International Patent Application Publication No. WO2014/116598, the entire contents of which are incorporated herein by reference, including all documents, tables, formulas, and figures. It is described in.

In this section, ratio comparisons or ratios or ratio values, ploidy estimates, and ploidy estimate values are collectively referred to as "comparisons." In some embodiments, the presence or absence of a chromosomal aneuploidy in a subject is determined according to one or more comparisons. In some embodiments, the presence or absence of a chromosomal aneuploidy in a subject is determined by one or more comparisons of three selected autosomes (e.g., one or more of three selected autosomes). is the test chromosome). In some embodiments, the presence or absence of chromosomal aneuploidy is generated for a series of different chromosomes, euploid regions, aneuploid regions, or euploid and aneuploid regions. determined according to one or more comparisons. In some embodiments, the presence or absence of chromosomal aneuploidy (e.g., chromosomal aneuploidy in a fetus) is determined by the comparison obtained for the subject and the euploid and/or aneuploid regions ( euploid region and aneuploid region determined for the reference set). In certain embodiments, the presence or absence of a chromosomal aneuploidy is determined according to the relationship between the comparison obtained for the subject and the euploid and/or aneuploid regions. For example, the presence or absence of a chromosomal aneuploidy, in some embodiments, whether the comparison is in a euploid region or an aneuploid region, or whether the ploidy assessment value is euploid Determined according to how far away it is from the region or aneuploid region. In some embodiments, the relationship is proximity or distance (eg, mathematical difference and/or graphical distance, eg, distance between points and regions). Relationships can be determined by any suitable method known in the art or described herein, non-limiting examples of which include probability distributions, probability density functions, cumulative distribution functions, likelihood functions , Bayesian model comparison, Bayes factor, deviance information criterion, chi-square test, Euclidean distance, spatial analysis, Mahalanobis distance, Manhattan distance, Chebyshev distance, Minkowski distance, Bregman divergence, Bhattacharya distance, Hellinger distance, metric space, Canberra Distances, convex hulls (eg, even-odd inflection rules), etc., or combinations thereof are included.

In some embodiments, the absence of chromosomal aneuploidy is determined according to comparison and euploid regions. In some embodiments, the absence of chromosomal aneuploidy is determined according to the relationship between the comparison and the euploid region. In some embodiments, the comparison within, in, or near the euploid region is a determination of a euploid chromosome (eg, the absence of an aneuploid chromosome). In some embodiments, a comparison that is in or near an euploid region indicates that each chromosome for which the comparison is determined is euploid. For example, sometimes comparisons generated according to counts mapped to ChrA, ChrB, and ChrC are determined according to counts mapped to euploid regions (e.g., ChrA, ChrB, and ChrC). euploid region) to determine the absence of chromosomal aneuploidy. In some embodiments, the absence of chromosomal aneuploidy is determined according to the comparison, where each chromosome (e.g., each chromosome from which a ploidy score is derived) is euploid (e.g., maternal and/or euploid in the fetus).

In some embodiments, the comparison outside the aneuploid region is determination of one or more euploid chromosomes. In some embodiments, a comparison that is outside the euploid region indicates that the chromosome or chromosomes for which the comparison was determined are euploid. For example, sometimes comparisons generated according to counts mapped to ChrA, ChrB, and ChrC are determined according to counts mapped to euploid regions (e.g., ChrA, ChrB, and ChrC). euploid region) to determine the absence of chromosomal aneuploidy. In some embodiments, comparisons that lie outside the euploid region are used for comparison or evaluation and indicate that two of the three chromosomes for which the comparison is determined are euploid.

In some embodiments, the comparison is within an aneuploid region and the one or more chromosomes for which the comparison is determined are euploid. For example, sometimes comparisons generated according to counts mapped to ChrA, ChrB, and ChrC are determined according to counts mapped to regions of aneuploids (e.g., ChrA, ChrB, and ChrC). aneuploid region) and the absence of chromosomal aneuploidy is determined for two of the three chromosomes.

In some embodiments, the presence of chromosomal aneuploidy is determined according to comparative and euploid regions. In certain embodiments, the presence of a chromosomal aneuploidy is determined according to the relationship between the comparison and the euploid region. In some embodiments, the comparison that is outside the euploid region is an aneuploid chromosome determination (eg, the presence of aneuploid chromosomes). In some embodiments, a comparison outside the euploid region indicates that the chromosome or chromosomes for which the comparison was determined are aneuploid. For example, sometimes comparisons generated according to counts mapped to ChrA, ChrB, and ChrC are determined according to counts mapped to euploid regions (e.g., ChrA, ChrB, and ChrC). euploid region) and determine the presence of chromosomal aneuploidy.

In some embodiments, the comparison within, in, or near an aneuploid region is an aneuploid chromosome determination (eg, the presence of an aneuploid chromosome). In some embodiments, a comparison that is in or near a region of aneuploid indicates that the one or more chromosomes for which the ploidy score was determined is aneuploid. In some embodiments, a comparison in or near a region of aneuploid indicates that the 1, 2, 3, 4, and/or 5 chromosomes for which the comparison was determined are aneuploid. . In some embodiments, a comparison in or near a region of aneuploid indicates that one of the three chromosomes for which the comparison was determined is aneuploid. For example, sometimes comparisons generated according to counts mapped to ChrA, ChrB, and ChrC are determined according to counts mapped to regions of aneuploids (e.g., ChrA, ChrB, and ChrC). aneuploid region) and one of the chromosomes is the aneuploid chromosome.

In some embodiments, the comparison near the aneuploid region is an aneuploid chromosome determination (eg, presence of an aneuploid chromosome). In some embodiments, a comparison near an aneuploid region indicates that the chromosome or chromosomes for which the comparison was determined are aneuploid. In some embodiments, the reference plot includes a defined euploid region and three defined aneuploid regions (e.g., aneuploid for Chr13, Chr18, or Chr21) and aneuploidy A determination of the presence of is made according to the closest comparison to one of the aneuploid regions. For example, a comparison that is closer to an aneuploid region for Chr21 than another region (e.g., an aneuploid region for Chr13 or Chr18, or a euploid region) indicates the presence of aneuploidy for Chr21. can point to

In some embodiments, comparisons generated according to counts mapped to Chr13, Chr18, and Chr21 are performed in aneuploid regions (e.g., counts mapped to Chr13, Chr18, and Chr21). aneuploid region determined according to ) and one of the chromosomes is an aneuploid chromosome. In some embodiments, comparisons generated according to counts mapped to Chr13, Chr18, and Chr21 are performed in aneuploid regions (e.g., counts mapped to Chr13, Chr18, and Chr21). Chr18 and Chr21 are determined to be euploid and Chr13 is determined to be aneuploid. In some embodiments, comparisons generated according to counts mapped to Chr13, Chr18, and Chr21 are performed in aneuploid regions (e.g., counts mapped to Chr13, Chr18, and Chr21). Chr13 and Chr21 are determined to be euploid and Chr18 is determined to be aneuploid. In some embodiments, comparisons generated according to counts mapped to Chr13, Chr18, and Chr21 are performed in aneuploid regions (e.g., counts mapped to Chr13, Chr18, and Chr21). Chr18 and Chr13 are determined to be euploid and Chr21 is determined to be aneuploid.

In some embodiments, the presence or absence of chromosomal aneuploidy is determined according to the first comparison and the second comparison. Here both comparisons were generated from sequence reads that mapped to the same set of two or more chromosomes. In some embodiments, the presence or absence of a chromosomal aneuploidy in a subject is determined by the relationship (e.g., distance) between a first comparison generated for the subject and a second comparison generated for the second subject. determined according to In some embodiments, the second comparison is a series of comparisons (eg, regions) generated for one or more subjects. In some embodiments, the presence or absence of a chromosomal aneuploidy in a subject is determined in relation to a first comparison generated for the subject and a reference set of comparisons generated for one or more subjects (e.g. , distance). In some embodiments, the first comparison is for a subject and the second comparison is a comparison or series of comparisons displaying one or more euploid fetuses. In some embodiments, the second comparison is an expected value or set of values (eg, area) for a euploid fetus. In some embodiments, the second comparison is generated for a subject (e.g., a pregnant female subject) whose fetus is known to be euploid for one or more of the chromosomes for which the comparison was generated. value or set of values. In some embodiments, distance is determined according to an uncertainty value (eg, standard deviation or MAD). In some embodiments, the distance between the first comparison and the second comparison (e.g., the second comparison displaying one or more euploid subjects) has an associated uncertainty of 1, 2, 3, 4, 5, 6 times or more and the first comparison is determined to be aneuploid. In some embodiments, the distance between the first comparison and the second comparison (e.g., the second comparison displaying one or more euploid subjects) is three times the associated uncertainty , or more, and the first comparison is determined to display an aneuploid chromosome.

In some embodiments, the presence or absence of chromosomal aneuploids is compared generated according to the number of counts mapped to one or more particular chromosomes and euploid regions, aneuploids or according to euploid and aneuploid regions. In some embodiments, the presence or absence of a chromosomal aneuploid is determined according to comparisons generated according to sequence reads mapped to one or more specific chromosomes, and to other chromosomes. Mapped sequence reads are not required for determination. In some embodiments, the presence or absence of a chromosomal aneuploid is determined according to comparisons generated according to reads of sequences mapped to 2, 3, 4, 5, or 6 different chromosomes; The number of counts mapped to the chromosome of is not obtained or required for determination. In some embodiments, the presence or absence of a chromosomal aneuploid is determined according to comparisons generated according to three different chromosomes or segments thereof, wherein the determination is a chromosome other than one of the three different chromosomes. not based on For example, if ChrA, ChrB, and ChrC represent three different chromosomes or segments thereof, the presence or absence of chromosomal aneuploids is sometimes determined according to comparisons made according to ChrA, ChrB, and ChrC, Decisions are not based on chromosomes other than ChrA, ChrB, or ChrC. In some embodiments, ChrA, ChrB, and ChrC represent Chr13, Chr21, and Chr18, respectively.

Sex Chromosome Karyotype In some embodiments, a principal component normalization process is used in conjunction with a method for determining the sex chromosome karyotype. Methods for determining sex chromosome karyotypes are disclosed, for example, in International Patent Application Publication No. WO2013/192562, the entire contents of which are incorporated herein by reference, including all documents, tables, formulas, and drawings. No.

In some embodiments, the read counts of sequences that map to one or more sex chromosomes (ie, chromosome X, chromosome Y) are normalized. In some embodiments, normalization includes principal component normalization. In some embodiments, normalization determines experimental bias for portions of the reference genome. In some embodiments, the experimental bias is the count of mapped sequence reads for each of the portions of the reference genome and the mapping feature (e.g., GC content) for each of the portions for each sample. A plurality of samples can be determined from a first fitted relationship (eg, a fitted linear relationship, a fitted non-linear relationship) for . The slope of the fitted relationship (eg, linear relationship) is generally determined by linear regression. In some embodiments, each experimental bias is indicated by an experimental bias coefficient. The experimental bias coefficient is, for example, the slope of the linear relationship between (i) the read counts of the sequence mapped to each of the parts of the reference genome and (ii) the mapping features for each of the parts. . In some embodiments, the experimental bias may include an estimate of the curvature of the experimental bias.

In some embodiments, the method includes a second fitted relationship (e.g., fitted linear relationship, fitted The step of calculating a level of genomic division (eg, elevation, level) for each of the genome portions from the non-linear relationship derived from the relationship), wherein the slope of the relationship can be determined by linear regression. For example, if the first fitted relationship is linear and the second fitted relationship is linear, the level of genome division Li can be determined for each of the parts of the reference _genome according to the formula α. :
L _i =(m _i −G _i S)I ⁻¹ Formula α

where G _i is the experimental bias, I is the intercept of the second fitted relationship, S is the slope of the second relationship, and _mi is each part of the reference genome. is the number of measurement counts mapped to , where i is the sample.

In some embodiments, a secondary normalization process is applied to the level of one or more calculated genome partitions. In some embodiments, secondary normalization includes GC normalization and sometimes the use of the PERUN method. In some embodiments, secondary normalization includes principal component normalization.

Determining Fetal Ploidy In some embodiments, the principal component normalization process is used in conjunction with a method for determining fetal ploidy. Methods for determining fetal ploidy are described, for example, in US Patent Application Publication No. 2013/0288244, the entire contents of which are incorporated herein by reference, including all documents, tables, formulas, and drawings. No.

Fetal ploidy can be determined in part from a measure of fetal fraction, which determines the presence or absence of genetic mutations (e.g., chromosomal aneuploidy, trisomy). used for Fetal ploidy can be determined in part from a measure of fetal fraction determined by any suitable method of determining fetal fraction, including the methods described herein. In some embodiments, the method calls for calculated reference counts F _i (sometimes also denoted as f _i ) determined for portions of the genome (i.e., bin i) for a plurality of samples. , where the polyploidy of the fetus for part i of the genome is known to be euploid. In some embodiments, an uncertainty value (eg, standard deviation, σ) is determined for the reference counts _fi . In some embodiments, reference counts f _i , uncertainty values, test sample counts and/or measured fetal fraction (F) are used to determine fetal ploidy. In some embodiments, reference counts (e.g., reference counts by mean, mean, or median) are subjected to principal component normalization and/or other normalizations, e.g., binwise normalization, GC content Normalize by normalization, linear least squares regression and non-linear least squares regression, LOESS, GC LOESS, LOWESS, PERUN, RM, GCRM, and/or combinations thereof. In some embodiments, a segment of the genome known to be euploid has a reference count equal to 1 when the reference count is normalized by principal component normalization. In some embodiments, both reference counts for portions or segments of the genome (e.g., for fetuses known to be euploid) and test sample counts are normalized by principal component normalization. , the reference count equals one. In some embodiments, a segment of the genome known to be euploid has a reference count equal to 1 when the reference count is normalized by PERUN. In some embodiments, both the reference counts for a portion or segment of the genome (e.g., for fetuses known to be euploid) and the test sample counts are normalized by PERUN, and the reference counts are number is equal to one. Similarly, in some embodiments, if the counts are normalized by the median reference count (i.e., divided by the median reference count), then genomes known to be euploid The reference count number of a part or segment of is equal to one. For example, in some embodiments, both the reference count (e.g., for fetuses known to be euploid) and the test sample count for a portion or segment of the genome are referred to as the reference count median. Normalized by value, the normalized reference count equals 1, and the test sample count is normalized by the median reference count (eg, divided by the median reference count). In some embodiments, both reference counts (e.g., for fetuses known to be euploid) and test sample counts for portions or segments of the genome are subjected to principal component normalization, GCRM , GC, RM, or any suitable method. In some embodiments, the reference count is a mean, average, or median reference count. Reference counts are often normalized counts for bins (eg, levels of normalized genomic divisions). In some embodiments, the reference counts and the counts for the test sample are raw counts. In some embodiments, the reference counts are determined from the mean, average, or median counts profile. In some embodiments, the reference count number is a calculated level of genome division. In some embodiments, reference counts of reference samples and counts of test samples (eg, patient samples, eg, y _i ) are normalized by the same method or process.

Additional Data Processing and Normalization Mapped sequence reads that have come to be counted are referred to herein as raw data, because these data contain unmanipulated counts (e.g. , unprocessed counts). In some embodiments, sequence read data in a dataset is further processed (e.g., mathematically and/or statistically manipulated) and/or displayed to facilitate obtaining an outcome. be able to. In certain embodiments, datasets, including larger datasets, may benefit from pre-processing to facilitate further analysis. Pre-processing of datasets sometimes includes overlapping and/or non-informative portions or portions of the reference genome (e.g., portions of the reference genome with non-informative data, overlapping, mapped reads, counts, etc.). including the removal of parts with a zero median, overrepresented or underrepresented sequences). Without being limited by theory, the processing and/or preprocessing of the data may include (i) removing noisy data, (ii) removing uninformative data, and (iii) removing redundant data. , (iv) reduce the complexity of larger data sets, and/or (v) facilitate the conversion of data from one form to one or more other forms. As used herein, the terms "pre-processing" and "processing" are collectively referred to as "processing" when used in relation to data or data sets. Processing can make the data more suitable for further analysis and, in some embodiments, can yield an outcome. In some embodiments, one or more or all of the processing methods (e.g., methods of normalization, partial filtering, mapping, validation, etc., or combinations thereof) are implemented in a processor, microprocessor, It is performed by a computer and/or by a microprocessor controlled device.

The term “noisy data,” as used herein, refers to (a) data that, when analyzed or plotted, show significant variance between data points; (b) have a significant standard deviation (e.g., 3 (c) data with a significant standard error of the mean, etc., and combinations of the above. Noisy data sometimes arises due to the quantity and/or quality of the starting material (e.g., nucleic acid sample) and sometimes is part of the process for preparing or replicating the DNA used to obtain sequence reads. generated from In certain embodiments, the noise results from overrepresentation of certain sequences when prepared using PCR-based methods. The methods described herein can reduce or eliminate the contribution of noisy data, thus reducing the effect of noisy data on the outcomes obtained.

The terms "non-informative data", "non-informative portion of the reference genome", and "non-informative portion", as used herein, are numerical values significantly different from a predetermined threshold value. , or a portion having a numerical value that lies outside a given range of values, or data derived therefrom. The terms "threshold" and "threshold value", as used herein, refer to any number calculated using a qualifying data set to determine genetic variation (e.g., copy number variation, aneuploidy, microscopic duplications, microdeletions, chromosomal aberrations, etc.). In certain embodiments, the results obtained by the methods described herein exceed a threshold and the subject is diagnosed with a genetic mutation (eg, trisomy 21). In some embodiments, the threshold value or range of values is often calculated by mathematically and/or statistically manipulating sequence read data (e.g., obtained from a reference and/or subject). and in certain embodiments, the sequence read data that is manipulated to obtain the threshold value or range of values is sequence read data (e.g., obtained from a reference and/or subject) . In some embodiments, an uncertainty value is determined. The uncertainty value is generally a measure of variance or error and may be any suitable measure of variance or error. In some embodiments, the uncertainty value is the standard deviation, standard error, calculated variance, p-value or mean absolute deviation (MAD). In some embodiments, the uncertainty value can be calculated according to the schemes described herein.

Any suitable procedure can be utilized to process the datasets described herein. Non-limiting examples of procedures suitable for use in processing data sets include filtering, normalizing, weighting, monitoring peak heights, and monitoring peak areas. monitoring edges of peaks; determining area ratios; mathematically processing data; statistically processing data; applying statistical algorithms; analysis with optimized variables, plotting data and identifying patterns or trends for further processing, etc., and combinations of the above. In some embodiments, various features (e.g., GC content, overlapping, mapped reads, centromeric regions, telomeric regions, etc., and combinations thereof) and/or variables (e.g., fetal sex, maternal Data sets are processed based on age of the mother, maternal ploidy, percentage contribution of fetal nucleic acid, etc., or a combination thereof). In certain embodiments, the complexity and/or dimensionality of large and/or complex datasets can be reduced by processing the datasets as described herein. Non-limiting examples of complex data sets include sequence read data generated from one or more test subjects and multiple reference subjects of different age and ethnic backgrounds. In some embodiments, a dataset can include thousands to millions of sequence reads for each test subject and/or reference subject.

In certain embodiments, data processing can occur in any number of steps. For example, in some embodiments, data can be processed using only a single processing procedure, and in certain embodiments, one or more, five or more, ten or more >, or 20 or more processing steps (e.g., 1 or more processing steps, 2 or more processing steps, 3 or more processing steps, 4 or more processing steps, 5 1 or more process steps, 6 or more process steps, 7 or more process steps, 8 or more process steps, 9 or more process steps, 10 or more process steps process steps, 11 or more process steps, 12 or more process steps, 13 or more process steps, 14 or more process steps, 15 or more process steps, 16 or more processing steps, 17 or more processing steps, 18 or more processing steps, 19 or more processing steps, or 20 or more processing steps) can be processed. In some embodiments, the processing step can be the same step repeated two or more times (e.g., filtering two or more times, normalizing two or more times), and certain In embodiments, the processing steps may be two or more different processing steps performed simultaneously or sequentially (e.g., filtering and normalizing; normalizing and monitoring peak heights and edges; filtering , normalized, normalized to the reference, statistically manipulated to determine the p-value, etc.). In some embodiments, any suitable number and/or combination of the same or different processing steps can be utilized to process sequence read data to facilitate obtaining an outcome. In certain embodiments, the complexity and/or dimensionality of the dataset can be reduced by processing the dataset according to the criteria described herein.

In some embodiments, one or more processing steps can include one or more filtering steps. The term "filtering" as used herein refers to removing portions or portions of a reference genome from consideration. Duplicate data (e.g., duplicate or overlapping mapped reads), non-informative data (e.g., portions of the reference genome with a median count of zero), overrepresented, including but not limited to Selecting and removing portions of the reference genome based on any suitable criteria, including portions of the reference genome with sequences that are overrepresented or underrepresented, noisy data, etc., or a combination of the above be able to. The filtering process often removes one or more portions of the reference genome from consideration and counts the number of counts in the one or more portions of the reference genome selected to be removed from the reference genome under consideration, one or more Including subtracting from counts counted or summed for multiple chromosomes or parts of the genome. In some embodiments, portions of the reference genome can be removed sequentially (e.g., removed one by one to allow assessment of the effects of removal of each individual portion), and certain In the embodiment of , all portions of the reference genome marked for removal can be removed at the same time. In some embodiments, portions of the reference genome characterized by variance above or below a certain level are removed, herein sometimes referred to as filtering "noisy" portions of the reference genome. call. In certain embodiments, the filtering process comprises obtaining data points from the data set that deviate from the mean profile level of the part, chromosome or chromosome segment by a predetermined multiple of the variance of the profile; In , the filtering process involves removing from the data set those data points that do not deviate from the mean profile level of the part, chromosome or chromosome segment by a predetermined multiple of the variance of the profile. In some embodiments, a filtering process is used to reduce the number of candidate parts of the reference genome that are analyzed for the presence or absence of genetic variation. Reduce the complexity and/or dimensionality of the dataset, often by reducing the number of candidate parts of the reference genome that are analyzed for the presence or absence of genetic mutations (e.g., microdeletion, microduplication) and sometimes increases the speed of searching and/or identifying genetic mutations and/or genetic aberrations by two orders of magnitude or more.

In some embodiments, one or more processing steps may include one or more normalization steps. Normalization can be performed by any suitable method described herein or known in the art. In certain embodiments, normalization includes adjusting values measured at different scales to a conceptually common scale. In certain embodiments, normalization involves sophisticated mathematical adjustments to bring a probability distribution of adjusted values into the alignment. In some embodiments, normalizing includes fitting a distribution to a normal distribution. In certain embodiments, normalization allows comparing normalized corresponding values for different data sets in a manner that eliminates the effects of certain global influences (e.g., errors and anomalies). contains a mathematical adjustment to make In certain embodiments, normalization includes scaling. Normalization sometimes involves dividing one or more data sets by a given variable or formula. Normalization sometimes involves subtraction of one or more data sets by a given variable or formula. Non-limiting examples of methods of normalization include normalization with respect to parts, normalization by GC content, normalization of median counts (median bin counts, median partial counts), linear and non-linear least squares Regression, LOESS, GC LOESS, LOWESS (Local Weighted Scatterplot Flattening), PERUN, ChAI, Principal Component Normalization, Repeat Masking (RM), GC Normalized Repeat Masking (GCRM), cQn, and/or Combinations thereof are included. In some embodiments, the determination of the presence or absence of a genetic mutation (e.g., aneuploidy, microduplication, microdeletion) is performed by a method of normalization (e.g., normalization for portions, normalization by GC content). normalization, median counts (median bin counts, median partial counts), linear and nonlinear least-squares regression, LOESS, GC LOESS, LOWESS (regional weighted scatterplot flattening), PERUN, ChAI , principal component normalization, repeat masking (RM), GC normalized repeat masking (GCRM), cQn, methods of normalization known in the art, and/or combinations thereof). In some embodiments, the determination of the presence or absence of a genetic mutation (e.g., aneuploidy, microduplication, microdeletion) is determined by LOESS, median counts (median bin counts, median fractional counts) value), and principal component normalization. In some embodiments, the determination of the presence or absence of a mutation in a gene utilizes LOESS followed by normalization of median counts (median bin counts, median fractional counts). In some embodiments, the determination of the presence or absence of a mutation in a gene utilizes LOESS followed by normalization of median counts (median bin counts, median fractional counts), Afterwards, principal component normalization is utilized.

Any suitable number of normalizations can be used. In some embodiments, the data set can be normalized one or more times, 5 or more times, 10 or more times, or even 20 or more times. Data sets can be normalized to values (eg, normalization values) that represent any suitable feature or variable (eg, sample data, reference data, or both). As a non-limiting example of the type of data normalization that can be used, raw count data for one or more selected test or reference moieties may be combined with the selected or normalizing to the total number of counts mapped to the chromosome to which the segment maps or to the entire genome; normalizing against the median reference counts for the one or more parts or chromosomes to which the processed parts or segments map; normalizing against derived values; and normalizing pre-normalized data against one or more other predetermined normalizing variables. Normalization of a data set sometimes has the effect of isolating statistical errors, depending on the feature or property chosen as the predetermined normalization variable. Normalization of data sets also sometimes allows comparison of data features of data having different scales by giving the data a common scale (eg, a predetermined normalization variable). In some embodiments, one or more normalizations to the statistically derived values can be utilized to minimize data differences and reduce the importance of outlier data. Normalizing a portion or portion of a reference genome with respect to a normalization value is sometimes referred to as "normalizing with respect to the portion."

In certain embodiments, the processing step including normalization includes normalizing to a stationary window; in some embodiments, the processing step including normalization includes normalizing to a moving or sliding window. including normalizing against The term "window," as used herein, refers to the portion or portions selected for analysis, sometimes used as a reference for comparison (e.g., normalization and/or other (used for mathematical or statistical manipulation of The term "normalize to a stationary window", as used herein, uses one or more portions selected to compare the data set under test and the data set under reference. It refers to the normalization process that In some embodiments, the selected portion is utilized to generate a profile. A static window generally includes a predetermined series of portions that do not change during manipulation and/or analysis. The terms "normalize against a moving window" and "normalize against a sliding window" as used herein refer to portions localized to genomic regions of the selected test portion (e.g., gene ), where one or more selected test portions are transformed to the nearest surrounding portion of the selected test portion. normalized to In one particular embodiment, the selected portion is utilized to generate a profile. Normalization of a sliding window or moving window is often repeatedly moved or slid toward adjacent test portions, and the newly selected test portion is placed in the immediate vicinity of the newly selected test portion or the newly selected test portion. normalizing to portions adjacent to the tested portion, where the adjacent windows have one or more portions in common. In certain embodiments, multiple selected test portions and/or chromosomes can be analyzed by sliding window processing.

In some embodiments, one or more values can be generated by normalizing against a sliding or moving window, where each value represents a different region of the genome (e.g. Display the results of normalization against a different set of reference parts selected from the chromosomes. In certain embodiments, the one or more values generated are a cumulative sum (e.g., of a selected portion, domain (e.g., portion of a chromosome) or integral of the count profile normalized across the chromosome). numerical estimates). Values generated by sliding window or moving window processing can be used to generate profiles and help reach outcomes. In some embodiments, the cumulative sum of one or more parts can be shown as a function of genomic position. Sometimes, moving window or sliding window analysis is used to analyze the genome for the presence or absence of microdeletions and/or microinsertions. In certain embodiments, representing the cumulative sum of one or more portions is used to identify the presence or absence of regions of genetic mutation (eg, microdeletion, microduplication). In some embodiments, moving window or sliding window analysis is used to identify genomic regions containing microdeletions, and in certain embodiments, moving window or sliding window analysis is used. to identify genomic regions containing microduplications.

Certain examples of normalization processes that can be utilized are described in more detail below, such as LOESS, PERUN, ChAI and principal component normalization methods.

In some embodiments, the processing step includes weighting. The terms “weighted,” “weighting,” or “weighting function,” or their grammatical derivatives or equivalents, as used herein, refer to the influence of a particular dataset feature or variable. , changes relative to other dataset features or variables (e.g., the significance and/or contribution of data contained in one or more portions or portions of the reference genome to the selected data set of the reference genome). Refers to the mathematical manipulation of part or all of a data set that may be utilized to increase or decrease based on the quality or usefulness of the data in one or more parts. In some embodiments, weighting functions are used to increase the influence of data with relatively small measurement variances and/or to decrease the influence of data with relatively large measurement variances. can be done. For example, portions of the reference genome that are underrepresented or have low quality sequence data can be "de-weighted" to minimize their impact on the dataset, while selected portions of the reference genome are "de-weighted". You can also "increase the weight" to increase the impact on the data set. A non-limiting example of a weighting function is [1/(standard deviation) ² ]. The weighting step is sometimes performed in a manner substantially similar to the normalization step. In some embodiments, the dataset is divided by a predetermined variable (eg, weighting variable). Often, a given variable (e.g., minimization objective function, Phi) is chosen to weight different parts of the data set differently (e.g., increase the influence of certain data types, while others reduce the impact of data type).

In certain embodiments, the processing step can include one or more mathematical and/or statistical manipulations. Any suitable mathematical and/or statistical manipulations can be used, alone or in combination, to analyze and/or manipulate the data sets described herein. Any suitable number of mathematical and/or statistical manipulations can be used. In some embodiments, the data set is mathematically and/or statistically manipulated one or more times, 5 or more times, 10 or more times, or 20 or more times be able to. Non-limiting examples of mathematical and statistical operations that can be used include addition, subtraction, multiplication, division, algebraic functions, least squares estimators, curve fitting, differential equations, rational polynomials, double polynomials, Orthogonal polynomials, z-scores, p-values, chi-values, phi-values, analysis of peak levels, determination of peak edge locations, calculation of peak area ratios, analysis of median values at chromosome level, calculation of mean absolute deviation, calculation of residuals Sum of squares, mean, standard deviation, standard error, etc., or combinations thereof. Mathematical and/or statistical manipulations can be performed on all or part of the sequence read data or their processed products. Non-limiting examples of dataset variables or features that can be statistically manipulated include raw counts, filtered counts, normalized counts, peak heights, peak widths, peak area, peak edge, lateral tolerance, P-value, median level, mean level, distribution of counts within a genomic region, relative representation of nucleic acid species, etc., or combinations thereof. .

In some embodiments, the processing step can involve the use of one or more statistical algorithms. Any suitable statistical algorithm can be used alone or in combination to analyze and/or manipulate the data sets described herein. Any suitable number of statistical algorithms can be used. In some embodiments, one or more, 5 or more, 10 or more, or 20 or more statistical algorithms can be used to analyze the data set. Non-limiting examples of statistical algorithms suitable for use with the methods described herein include decision trees, alternative hypotheses, multiple comparisons, omnibus tests, Behrens-Fischer tests, bootstrap methods, independent significance Fisher method for combining sex tests, null hypothesis, type I error, type II error, exact test, one-sample Z-test, two-sample Z-test, one-sample t-test, paired t-test, equal variances 2-sample pooled t-test with unequal variances, 2-sample unpooled t-test with unequal variances, 1-proportion z-test, 2-proportion z-test pooled, 2-proportion z-test unpooled, 1-sample chi-square test, one-of-variance Two-sample F-test for similarity, confidence intervals, credible intervals, significance, meta-analysis, simple linear regression, robust linear regression, etc., or combinations of the above. Non-limiting examples of dataset variables or features that can be analyzed using statistical algorithms include raw counts, filtered counts, normalized counts, peak heights, peak width of a peak, edge of a peak, lateral tolerance, P-value, median level, mean level, distribution of counts within a genomic region, relative representation of nucleic acid species, etc., or combinations thereof.

In certain embodiments, multiple (e.g., two or more) statistical algorithms (e.g., least squares regression, principal component analysis, linear discriminant analysis, quadratic discriminant analysis, bagging, neural networks, support vector machines models, random forests, classification tree models, K nearest neighbors, logistic regression and/or loss smoothing) and/or mathematical and/or statistical manipulations (e.g., referred to herein as manipulations). can analyze the dataset. In some embodiments, the use of multiple operations can generate an N-dimensional space that can be used to provide outcomes. In certain embodiments, the complexity and/or dimensionality of the dataset can be reduced by analyzing the dataset by utilizing multiple operations. For example, multiple manipulations can be used on a reference data set to determine the presence or absence of mutations in genes depending on the genetic context of the reference sample (e.g., positive or negative for mutations in selected genes). An N-dimensional space (eg, probability plot) can be generated that can be used for display. Analysis of the test samples using a series of substantially similar operations can be used to generate an N-dimensional point for each of the test samples. The complexity and/or dimensionality of the dataset under test is sometimes simplified to a single value or N-dimensional point that can be easily compared to the N-dimensional space generated from the reference data. Test sample data belonging to the N-dimensional space in which the referenced data resides exhibit a gene context that is substantially similar to the referenced gene context. Test sample data that resides outside the N-dimensional space in which the referenced data resides exhibit gene context that is substantially dissimilar to the referenced gene context. In some embodiments, the reference is euploid and otherwise has no genetic mutations or medical conditions.

In some embodiments, after the data sets have been counted, optionally filtered and normalized, these processed data sets by one or more steps of filtering and/or normalizing can be further manipulated. In certain embodiments, a profile can be generated using a data set that has been further manipulated by one or more filtering and/or normalizing procedures. In some embodiments, one or more filtering and/or normalizing procedures can sometimes reduce the complexity and/or dimensionality of the dataset. Outcomes can be provided based on reduced complexity and/or dimensionality data sets.

In some embodiments, a measure of error (e.g., standard deviation, standard error, calculated variance, p-value, mean absolute error (MAE), mean absolute deviation and/or mean absolute deviation (MAD )) according to which the part can be filtered. In certain embodiments, the error measure refers to the variability of counts. In some embodiments, the portions are filtered according to the variability of the counts. In certain embodiments, the variability in counts is measured in multiple samples (e.g., multiple subjects, e.g., 50/animal or more, 100/animal or more, 500/animal or more, 1000/mouse or more, 5000/mouse or more, or 10,000/mouse or more subjects), for a portion (i.e., portion) of the reference genome is a measure of the error determined for counts mapped by In some embodiments, portions with count variability above a predetermined upper range are filtered (eg, eliminated from consideration). In some embodiments, the predetermined upper range is equal to or greater than about 50, equal to or greater than about 52, equal to or greater than about 54, equal to or greater than about 56, equal to about 58 or greater than, equal to or greater than about 60, equal to or greater than about 62, equal to or greater than about 64, equal to or greater than about 66, equal to or greater than about 68, equal to or greater than about 70 , equal to or greater than about 72, equal to or greater than about 74, or equal to or greater than about 76. In some embodiments, portions with count variability below a predetermined lower range are filtered (eg, eliminated from consideration). In some embodiments, the predetermined lower range is less than or equal to about 40, less than or equal to about 35, less than or equal to about 30, less than or equal to about 25, less than or equal to about 20, or MAD values less than, less than or equal to about 15, less than or equal to about 10, less than or equal to about 5, less than or equal to about 1, or less than or equal to about 0. In some embodiments, portions with count variability outside a predetermined range are filtered (eg, eliminated from consideration). In some embodiments, the predetermined range is from greater than zero to less than about 76, less than about 74, less than about 73, less than about 72, less than about 71, less than about 70, less than about 69, less than about 68, about 67 A MAD value of less than, less than about 66, less than about 65, less than about 64, less than about 62, less than about 60, less than about 58, less than about 56, less than about 54, less than about 52, or less than about 50. In some embodiments, the predetermined range is a MAD value from greater than zero to less than about 67.7. In some embodiments, portions with count variability within a predetermined range are selected (eg, used to determine the presence or absence of mutations in a gene).

In some embodiments, the variability in the fractional counts exhibits a distribution (eg, a normal distribution). In some embodiments, the portion is selected within a quantile of the distribution. In some embodiments, less than or equal to about 99.9%, less than or equal to about 99.8%, less than or equal to about 99.7%, less than or equal to about 99.6% of the distribution, or less than, equal to or less than about 99.5%, equal to or less than about 99.4%, equal to or less than about 99.3%, equal to or less than about 99.2%, about 99.5%; Equal to or less than 1%, Equal to or less than about 99.0%, Equal to or less than about 98.9%, Equal to or less than about 98.8%, Equal to or less than about 98.7% less than, less than or equal to about 98.6%, less than or equal to about 98.5%, less than or equal to about 98.4%, less than or equal to about 98.3%, about 98.2 % equal to or less than, equal to or less than about 98.1%, equal to or less than about 98.0%, equal to or less than about 97%, equal to or less than about 96%, about 95% Equal to or less than, Equal to or less than about 94%, Equal to or less than about 93%, Equal to or less than about 92%, Equal to or less than about 91%, Equal to or less than about 90% , less than or equal to about 85%, less than or equal to about 80%, or less than or equal to about 75% within the quantile is selected. In some embodiments, the portion within the 99% quantile of the count variability distribution is selected. In some embodiments, within the 99% quantile, parts with MAD>0 and parts with MAD<67.725 are selected, resulting in the identification of a set of stable parts of the reference genome.

Non-limiting examples of filtering portions for PERUN are given, for example, herein and in International Patent Application No. PCT/US12/59123 (WO2013/052913), the latter for all documents, Its entire contents, including tables, formulas and figures, are incorporated herein by reference. Portions can be filtered based on the error measure or based in part on the error measure. In certain embodiments, a measure of error, including the absolute value of the deviation, such as the R-factor, can be used to remove or weight portions. The R-factor, in some embodiments, is defined as the sum of the absolute deviations of the count values expected from the actual measurements divided by the count values expected from the actual measurements. A measure of error that includes the absolute value of the deviation can be used, but any suitable measure of error can be used instead. In certain embodiments, a measure of error that does not include the absolute value of the deviation can be utilized, such as square-based variation. In some embodiments, portions are filtered or weighted according to a mappability measure (eg, a mappability score). Sometimes, the portion is filtered according to a relatively low number of sequence reads mapped to the portion (e.g., 0, 1, 2, 3, 4, 5 reads mapped to the portion). or weighted. Portions can be filtered or weighted according to the type of analysis being performed. For example, in the case of analysis of aneuploidy of chromosomes 13, 18 and/or 21, sex chromosomes can be filtered and only autosomes or a subset of autosomes can be analyzed.

Certain embodiments may utilize the following filtering process. The same set of segments (eg, segments of the reference genome) within a given chromosome (eg, chromosome 21) is selected and the numbers of reads are compared between diseased and non-diseased samples. Gaps relate trisomy 21 and euploid samples, and include a series of segments covering most of chromosome 21. These series of parts are the same between euploid and T21 samples. Since the parts can be defined, the distinction between a series of parts and a single partition is less important. The same genomic regions are compared in different patients. This treatment can be utilized for trisomy analysis, eg, T13 or T18, in addition to or instead of T21.

In some embodiments, these processed data sets can be manipulated by weighting after the data sets have been counted, optionally filtered and normalized. In certain embodiments, one or more portions are selected and weighted to account for the impact of data contained in the selected portions (e.g., noisy data, uninformative data). can be reduced, and in some embodiments, one or more portions are selected and weighted to determine the data contained in the selected portions (e.g., data for which a small variance was measured) can enhance or increase the effect of In some embodiments, the dataset is weighted using a single weighting function that reduces the impact of data with large variances and increases the impact of data with small variances. Sometimes a weighting function is used to reduce the influence of data with large variances and increase the influence of data with small variances (eg, [1/(standard deviation) ² ]). In some embodiments, weightings are further manipulated to produce profile plots of the processed data to facilitate classification and/or provision of outcomes. Outcomes can be generated based on plots of weighted data profiles.

Filtering or weighting the portions can be done at one or more suitable points in the analysis. For example, before or after mapping sequence reads to portions of a reference genome, portions can be filtered or weighted. In some embodiments, portions can be filtered or weighted before or after determining experimental bias for individual genomic portions. In certain embodiments, portions can be filtered or weighted before or after calculating levels of genome division.

In some embodiments, after the data sets have been counted, optionally filtered, normalized and optionally weighted, these processed data sets are subjected to one or more mathematical and/or or can be manipulated statistically (eg, by statistical functions or statistical algorithms). In certain embodiments, the processed dataset can be further manipulated by calculating Z-scores for one or more selected portions, chromosomes, or portions of chromosomes. In some embodiments, the processed dataset can be further manipulated by calculating P-values. In certain embodiments, the mathematical and/or statistical manipulation includes one or more assumptions regarding ploidy and/or fetal fraction. In some embodiments, the processed data is further manipulated by one or more statistical and/or mathematical manipulations to generate profile plots to facilitate classification and/or provision of outcomes . Outcomes can be generated based on profile plots of statistically and/or mathematically manipulated data. Outcomes derived from profile plots of statistically and/or mathematically manipulated data often include one or more assumptions regarding ploidy and/or fetal fraction.

In certain embodiments, after the data set has been counted, optionally filtered and normalized, multiple operations are performed on the processed data set to determine the N-dimensional space and/or N-dimensional generate points for . Outcomes can be generated based on plots of profiles of datasets analyzed in N dimensions.

In some embodiments, as part of or after processing and/or manipulating the data set, one or more of peak level analysis, peak width analysis, peak edge location analysis, peak lateral tolerance analysis, etc. etc., derivatives thereof, or combinations of the above, to process the data set. In some embodiments, processing utilizes one or more of peak level analysis, peak width analysis, peak edge location analysis, peak lateral tolerance, etc., derivatives thereof, or combinations of the above. Produce profile plots of collected data to facilitate classification and/or provision of outcomes. profile of data that has been processed using one or more of peak level analysis, peak width analysis, peak edge location analysis, peak lateral tolerance, etc., derivatives thereof, or combinations of the above; Outcomes can be delivered based on plots.

In some embodiments, one or more reference samples that are substantially free of mutations in the gene of interest can be used to obtain a median reference count profile, which is a profile of mutations in the gene. often, if the test subject carries a mutation in the gene, from a predetermined value in the region corresponding to the genomic location where the mutation in the gene is located in the test subject. Profile deviates. In test subjects at risk for, or suffering from, a medical condition associated with a genetic mutation, values for selected segments or segments are compared to their genomic locations when not affected. It is expected to be significantly different from the prescribed value. In certain embodiments, one or more reference samples known to carry mutations in the gene of interest can be used to obtain a median reference count profile, which is a profile of the gene The profile deviates from a predetermined value that can be indicative of the presence of a mutation, and often in regions corresponding to locations in the genome where the test subject does not carry a mutation in that gene. In test subjects who are not at risk for or do not suffer from medical conditions associated with genetic mutations, values for selected segments or segments are is expected to be significantly different from the predetermined value of .

In some embodiments, analysis and processing of data may involve the use of one or more assumptions. A suitable number or type of assumptions can be utilized to analyze or process the data set. Non-limiting examples of assumptions that can be used for data processing and/or analysis include maternal ploidy, fetal contribution, prevalence of a particular sequence in a reference population, ethnic background, consanguinity prevalence of selected medical conditions in families with different patients, degree of parallelism between profiles of raw counts obtained from different patients and/or runs after GC-normalized repeat masking (e.g., GCRM), lack of PCR. Identical matches implying natural outcomes (e.g. identical base positions), assumptions inherent in fetal quantification assays (e.g. FQA), assumptions about twins (e.g. both twins, only one affected) If so, the effective fetal fraction is only 50% of the total measured fetal fraction (also for triplets, quadruplets, etc.), and fetal cell-free DNA that uniformly covers the entire genome (e.g. , cfDNA), etc., as well as combinations thereof.

Predicting the outcome of the presence or absence of a genetic mutation with a desired level of confidence (e.g., a level of confidence of 95% or greater) based on the normalized count profile is mapped. Utilizing one or more additional mathematical manipulation algorithms and/or statistical prediction algorithms to analyze data and/or provide outcomes in cases where the read quality and/or depth of the sequence read is not possible can generate additional numbers useful for The term "normalized count profile" as used herein refers to a profile generated using normalized counts. Examples of methods that can be used to generate normalized counts and normalized count profiles are described herein. As noted above, the mapped sequence reads that come to be counted can be normalized with respect to the test sample counts or the reference sample counts. In some embodiments, the normalized count profile can be plotted and shown.

LOESS Regularization LOESS is a regression modeling method known in the art that combines multiple regression models within a k-nearest neighbor-based metamodel. LOESS is sometimes referred to as locally weighted polynomial regression. In some embodiments, GC LOESS applies a LOESS model to the relationship between fragment counts (eg, sequence reads, sequence counts) and GC composition for a reference genome portion. Plotting a smooth curve through a set of data points using LOESS is sometimes referred to as a LOESS curve, specifically where each smoothed value is the value of the scatterplot reference variable on the y-axis. It is so called if it is given by a weighted quadratic least squares regression over intervals. For each point in the data set, the LOESS method fits a low-order polynomial to the subset of data whose explanatory variable values are near the point whose response is estimated. The polynomial is fit using a weighted least squares method that gives greater weight to points near the point whose response is estimated and less weight to points farther away. A regression function value for a point is then obtained by evaluating the local polynomial using the explanatory variable values for that data point. A LOESS fit is sometimes considered complete after a regression function value has been calculated for each of the data points. Many of the details of the method, such as the polynomial model order and weights, are adaptive.

PERUN normalization Normalization methods for reducing errors associated with nucleic acid signatures are referred to herein as PERUN (parameterized error removal and unbiased normalization), which are described herein as well as in the text, tables, formulas, and in International Patent Application Publication No. WO2013/052913, the entire contents of which are incorporated herein by reference, including all of the drawings. The PERUN method can be applied to a variety of nucleic acid beacons (eg, nucleic acid sequence reads) to reduce the effects of errors that confound predictions based on such indices.

In certain embodiments, the PERUN method calculates the level of genomic segmentation for a reference genome portion as: (a) for the test sample, the count of reads of sequences that mapped to the reference genome portion; , the experimental bias (e.g., GC bias), and (c) (i) the experimental bias for the reference genome portion to which the sequence reads map and (ii) the sequence reads mapped to the portion. calculating from one or more fitted parameters (eg, an estimate of the fit) the fitted relationship between the counts of . The experimental bias for each of the reference genome portions is a fitted relationship for each sample across multiple samples comprising: (i) the read count of the sequence that mapped to each of the reference genome portions; (ii) can be determined according to the relationship with the mapping features for each of the reference genome portions; This fitted relationship for each sample can be assembled in three dimensions for multiple samples. In certain embodiments, assembly can be ordered according to experimental bias, but the PERUN method can also be performed without ordering assembly according to experimental bias. The fitted relationship for each sample and the fitted relationship for each portion of the reference genome can be independently fit to a linear or non-linear function by any suitable fitting process known in the art. can be done.

Hybrid Regularization for Regression In some embodiments, a hybrid regularization method is used. In some embodiments, hybrid normalization methods reduce bias (eg, GC bias). In some embodiments, hybrid normalization consists of (i) an analysis of the relationship between two variables (e.g., counts and GC content) and (ii) selection and application of a normalization method according to the analysis. including. In certain embodiments, hybrid normalization includes (i) regression (eg, regression analysis) and (ii) selection and application of a normalization method according to the regression. In some embodiments, the counts obtained for a first sample (e.g., first set of samples) are determined in a different manner than the counts obtained from another sample (e.g., second set of samples). Normalize. In some embodiments, counts obtained for a first sample (e.g., first sample set) are normalized by a first normalization method to obtain a second sample (e.g., second sample set) ) are normalized by a second normalization method. For example, in certain embodiments, the first normalization method includes using linear regression and the second normalization method includes using non-linear regression (e.g., LOESS, GC-LOESS, LOWESS regression, LOESS smoothing). Including use.

In some embodiments, hybrid normalization methods are used to normalize sequence reads (eg, counts, mapped counts, mapped reads) to portions of a genome or chromosome. In certain embodiments, raw counts are normalized, and in some embodiments, adjusted, weighted, filtered, or already normalized counts are subjected to hybrid normalization. Normalize by the normalization method. In certain embodiments, the levels of genomic divisions or Z-scores are normalized. In some embodiments, the counts that map to the selected genome portion or chromosome are normalized by hybrid normalization methods. Counts are a suitable measure of sequence reads that map to a portion of the genome, non-limiting examples of which are raw counts (e.g., unprocessed counts), normalized Counts (e.g., PERUN, ChAI, principal component normalization, or normalized by an appropriate method), partial levels (e.g., mean level, mean level, median level, etc.), z-scores, etc., or any of these Sometimes refers to a scale that includes combinations. The counts can be raw counts or processed counts from one or more samples (eg, test samples, samples from pregnant females). In some embodiments, counts are obtained from one or more samples obtained from one or more subjects.

In some embodiments, the normalization method (eg, type of normalization method) is selected according to regression (eg, regression analysis) and/or correlation coefficients. Regression analysis refers to a statistical technique for estimating relationships between variables (eg counts and GC content). In some embodiments, regressions are generated according to counts and measures of GC content for each of the plurality of portions of the reference genome. Suitable measures of GC content, non-limiting examples of which are guanine content, cytosine content, adenine content, thymine content, purine (GC) content, or pyrimidine (AT or ATU) content Measures including quantity measures, melting temperatures (T _m ) (eg, denaturation temperature, annealing temperature, hybridization temperature), free energy measures, etc., or combinations thereof, can be used. A measure of guanine (G) content, cytosine (C) content, adenine (A) content, thymine (T) content, purine (GC) content, or pyrimidine (AT or ATU) content is the ratio or It can be expressed as a percentage. In some embodiments, any suitable ratio or percentage, non-limiting examples of which are GC/AT, GC/total nucleotides, GC/A, GC/T, AT/total nucleotides, AT/GC , AT/G, AT/C, G/A, C/A, G/T, G/A, G/AT, C/T, etc., or combinations thereof. In some embodiments, the measure of GC content is the ratio or percentage of GC content to total nucleotide content. In some embodiments, the measure of GC content is the ratio or percentage of GC content to total nucleotide content for sequence reads that map to a portion of the reference genome. In certain embodiments, the GC content is determined according to and/or from sequence reads mapped to each reference genome portion, wherein the sequence reads are determined from the sample ( For example, samples obtained from pregnant females). In some embodiments, the measure of GC content is not determined according to and/or from the sequence reads. In certain embodiments, a measure of GC content is determined for one or more samples obtained from one or more subjects.

In some embodiments, generating a regression includes generating a regression analysis or a correlation analysis. Non-limiting examples are regression analysis (e.g., linear regression analysis), goodness-of-fit analysis, Pearson correlation analysis, rank correlation, proportion of unexplained variance, NS (Nash-Sutcliffe) model efficiency Any suitable regression can be used, including analysis, validation of regression models, proportional reduction in loss (PRL), root mean square deviation, etc., or a combination thereof. In some embodiments, a regression line is generated. In certain embodiments, generating a regression includes generating a linear regression. In certain embodiments, generating a regression includes generating a non-linear regression (eg, LOESS regression, LOWESS regression).

In some embodiments, regression determines, for example, the presence or absence of a correlation (eg, a linear correlation) between counts and measures of GC content. In some embodiments, a regression (eg, linear regression) is generated and the correlation coefficient determined. In some embodiments, a suitable correlation coefficient is determined, non-limiting examples of which include the coefficient of determination, the ^R2 value, the Pearson correlation coefficient, and the like.

In some embodiments, goodness of fit is determined for regression (eg, regression analysis, linear regression). Goodness of fit is optionally determined by visual or mathematical analysis. Evaluation optionally includes determining whether the goodness of fit is greater for nonlinear regression or linear regression. In some embodiments, the correlation coefficient is a measure of goodness of fit. In some embodiments, the goodness-of-fit rating for the regression is determined according to the correlation coefficient and/or the correlation coefficient cutoff value. In some embodiments, evaluating goodness of fit includes comparing the correlation coefficient to a cutoff value for the correlation coefficient. In some embodiments, the goodness-of-fit evaluation for regression is indicative of linear regression. For example, in certain embodiments, the goodness of fit is greater for linear regression than for nonlinear regression, and the goodness of fit evaluation is indicative of linear regression. In some embodiments, evaluation refers to linear regression, which is used to normalize counts. In some embodiments, the goodness-of-fit evaluation for regression is indicative of non-linear regression. For example, in certain embodiments, the goodness of fit is greater for nonlinear regression than for linear regression, and the goodness of fit evaluation is indicative of nonlinear regression. In some embodiments, evaluation refers to non-linear regression, which is used to normalize counts.

In some embodiments, an assessment of goodness of fit indicates linear regression when the correlation coefficient is equal to or greater than the correlation coefficient cutoff. In some embodiments, the goodness-of-fit evaluation indicates non-linear regression if the correlation coefficient is less than the correlation coefficient cutoff. In some embodiments, the correlation coefficient cutoff is a predetermined cutoff. In some embodiments, the correlation coefficient cutoff is about 0.5 or greater, about 0.55 or greater, about 0.6 or greater, about 0.65 or greater, about 0.7 or greater, about 0.75 or greater, about 0.8 or greater, or about 0.85 or greater.

For example, in one particular embodiment, normalization methods involving linear regression are used when the correlation coefficient is equal to or greater than about 0.6. In certain embodiments, if the correlation coefficient is equal to or greater than a correlation coefficient cutoff of 0.6, the number of counts per sample (e.g., reference genome portion, counts) are normalized according to linear regression, otherwise counts are normalized according to non-linear regression (eg, if the coefficient is less than the correlation coefficient cutoff of 0.6). In some embodiments, the normalization process comprises generating a linear or non-linear regression for (i) counts and (ii) GC content, each portion of the plurality of portions of the reference genome. . In certain embodiments, normalization methods including nonlinear regression (eg, LOWESS, LOESS) are used when the correlation coefficient is below a correlation coefficient cutoff of 0.6. In some embodiments, a correlation with a correlation coefficient (e.g., correlation coefficient) of about 0.7, less than about 0.65, less than about 0.6, less than about 0.55, or less than about 0.5 If less than a number cutoff, use regularization methods including non-linear regression (eg LOWESS). For example, some embodiments use normalization methods including nonlinear regression (eg, LOWESS, LOESS) when the correlation coefficient is below a correlation coefficient cutoff of about 0.6.

In some embodiments, after selecting a specific type of regression (eg, linear or non-linear regression) and generating the regression, the counts are normalized by subtracting the regression from the counts. In some embodiments, the regression is subtracted from the counts to present normalized counts with reduced bias (eg, GC bias). In some embodiments, linear regression is subtracted from counts. In some embodiments, a non-linear regression (eg, LOESS, GC-LOESS, LOWESS regression) is subtracted from the counts. Any suitable method can be used to subtract the regression line from the counts. For example, the number of counts x is derived from the portion i containing a GC content of 0.5, and the regression line determines the number of counts y when the GC content is 0.5, thus x−y= is the normalized count number for portion i. In some embodiments, the counts are normalized before subtracting the regression and/or after subtracting the regression. In some embodiments, hybrid normalization method normalized counts are used to generate levels of genome divisions, Z-cores, levels and/or profiles of genomes or segments thereof. In certain embodiments, hybrid normalization method normalized counts are analyzed by methods described herein to determine the presence or absence of a genetic mutation (e.g., in a fetus). .

In some embodiments, hybrid normalization methods include filtering or weighting one or more portions before or after normalization. Any suitable portion filtering method can be used, including the portion (eg, reference genome portion) filtering methods described herein. In some embodiments, portions (eg, reference genome portions) are filtered prior to applying the hybrid normalization method. In some embodiments, hybrid normalization normalizes only the counts of sequencing reads that map to selected portions (eg, portions selected according to count variability). In some embodiments, exclude counts of sequencing reads that mapped to filtered reference genome portions (e.g., portions filtered according to count variability) prior to leveraging hybrid normalization methods. . In some embodiments, hybrid normalization methods involve selecting or filtering portions (e.g., reference genome portions) according to suitable methods (e.g., methods described herein). . In some embodiments, the hybrid normalization method selects or filters portions (e.g., reference genome portions) according to uncertain values for the number of counts mapped to each of the portions for multiple test samples. including doing In some embodiments, hybrid normalization methods include selecting or filtering portions (eg, reference genome portions) according to count variability. In some embodiments, the hybrid normalization method selects or filters portions (e.g., reference genome portions) according to GC content, repetitive elements, repetitive sequences, introns, exons, etc., or combinations thereof. including doing

For example, in some embodiments, multiple samples from multiple pregnant female subjects are analyzed and subsets of portions (eg, reference genome portions) are selected according to the variability in counts. In certain embodiments, linear regression is used to calculate correlation coefficients for (i) counts and (ii) GC content for each of the selected portions for samples obtained from pregnant female subjects. decide. In some embodiments, a correlation coefficient exceeding a predetermined correlation cutoff value (e.g., a correlation cutoff value of about 0.6) is determined, and the goodness-of-fit evaluation indicates linear regression; Normalize the counts by subtracting from the counts. In certain embodiments, correlation coefficients below a predetermined correlation cutoff value (e.g., a correlation cutoff value of about 0.6) are determined, and goodness-of-fit assessments indicate nonlinear regression and LOESS regression. Generate and normalize the counts by subtracting the LOESS regression from the counts.

Profile In some embodiments, the step of processing comprises various aspects of the dataset or derivative thereof (e.g., one or more profiles known in the art and/or described herein). generating one or more profiles (eg, plotting the profile) from the result of a mathematical data processing step and/or a statistical data processing step).

As used herein, the term "profile" refers to the result of mathematical and/or statistical manipulation of data that can facilitate identification of patterns and/or correlations in large amounts of data. A "profile" often includes values resulting from one or more operations on data or data sets based on one or more reference criteria. Profiles often contain multiple data points. Any suitable number of data points can be incorporated into the profile, depending on the nature and/or complexity of the dataset. In certain embodiments, the profile includes 2 or more data points, 3 or more data points, 5 or more data points, 10 or more data points, 24 or more data points, 25 or more data points, 50 or more data points, 100 or more data points, 500 or more data points, 1000 or more data points, 5000 or more data points points, 10,000 or more data points, or 100,000 or more data points can be incorporated.

In some embodiments, the profile displays the entire dataset, while in certain embodiments the profile displays a portion or subset of the dataset. That is, in some cases the profile includes or is generated from data points representing data that has not been filtered to exclude any data, and in some cases the profile contains undesirable data. Contains or is generated from data points representing data that has been filtered to exclude. In some embodiments, the data points in the profile display the results of data manipulation for the portion. In certain embodiments, the data points in the profile include the results of data manipulation for groups of parts. In some embodiments, groups of portions can be adjacent to each other, and in certain embodiments, groups of portions can be derived from different parts of the chromosome or genome.

The data points in the profile derived from the dataset may represent any suitable data categorization. Non-limiting examples of categories by which data can be grouped to generate profile data points include segments based on size, sequence features (e.g., GC content, AT content, chromosomal locations (e.g., short arm segment, long arm segment, centromere, telomere), level of expression, chromosome, etc., or combinations thereof. In some embodiments, a profile can be generated from data points obtained from another profile (e.g., renormalized data points renormalized according to different normalization values to generate a renormalized data profile). data profile). In certain embodiments, a profile generated from data points obtained from another profile reduces the number of data points and/or the complexity of the data set. Reducing the number of data points and/or the complexity of the dataset often facilitates data interpretation and/or presentation of outcomes.

A profile (e.g., genomic profile, chromosomal profile, segmental profile of a chromosome) is often a collection of normalized or non-normalized counts of two or more parts. A profile often includes at least one level (eg, the level of a genome division) and often includes two or more levels (eg, a profile often has multiple levels). A level is generally a level for a set of parts that have approximately the same count or normalized count. Levels are described in more detail herein. In certain embodiments, a profile is one or more parts that are weighted, excluded, filtered, normalized, adjusted, averaged, or derived as an average. , which may be added, subtracted, manipulated, or transformed by any combination thereof. Profiles often include normalized counts mapped to portions defining two or more levels, where the counts are further mapped according to one of the levels by a suitable method. Normalized. Profile (eg, profile level) counts are often associated with uncertain values.

Profiles that include one or more levels are sometimes filled (eg, filled with holes). Hole-filling (eg, hole-filling) refers to the process of identifying and adjusting levels in a profile due to maternal microdeletion or duplication (eg, copy number variation). In some embodiments, levels resulting from fetal microduplication or fetal microdeletion are bridged. In some embodiments, micro-duplications or micro-deletions in the profile artificially increase or decrease the overall level of the profile (e.g., chromosomal profile), and for chromosomal aneuploidy (e.g., trisomy) , can result in false positive or false negative determinations. In some embodiments, levels in the profile due to microduplications and/or deletions are identified and optionally adjusted (eg, filled and/or excluded) through a process referred to as filling holes or filling holes. In certain embodiments, the profile comprises one or more first levels that are significantly different from a second level in the profile, each of the one or more first levels modulating one or more of the first levels, including number variation, fetal copy number variation, or maternal and fetal copy number variation.

A profile that includes one or more levels can include a first level and a second level. In some embodiments, the first level is different (eg, significantly different) than the second level. In some embodiments, the first level includes a first set of portions, the second level includes a second set of portions, and the first set of portions includes a second set of portions. not a subset of the set. In certain embodiments, the first set of portions is different from the second set of portions from which the first level and the second level are determined. In some embodiments, a profile may have multiple first levels that are different (eg, significantly different, eg, have significantly different values) than second levels in the profile. In some embodiments, the profile includes one or more first levels that are significantly different from a second level in the profile, and one or more of the first levels are adjusted. In some embodiments, the profile comprises one or more first levels that are significantly different from a second level in the profile, each of the one or more first levels modulating one or more of the first levels, including number variation, fetal copy number variation, or maternal and fetal copy number variation. In some embodiments, the first level in the profile is excluded from the profile or adjusted (eg, filled in). A profile can include multiple levels, including one or more first levels that are significantly different from one or more second levels, wherein most of the levels in the profile are substantially similar to each other. Often an equal second level. In some embodiments, more than 50%, 60%, 70%, 80%, 90% or 95% of the levels in the profile are second levels.

Profiles are sometimes shown as plots. For example, one or more levels representing partial counts (eg, normalized counts) can be plotted and visualized. Non-limiting examples of plots of profiles that can be generated include raw counts (e.g., raw counts profile or unprocessed profile), normalized counts, partial weights, z-scores, p-values , area ratio versus fitted ploidy, median level versus ratio of fitted and measured fetal fractions, principal components, etc., or combinations thereof. In some embodiments, profile plots allow visualization of operational data. In certain embodiments, profile plots are utilized to determine outcomes (e.g., area ratio versus fitted ploidy, median level versus ratio of fitted to measured fetal fraction, main ingredients) can be presented. As used herein, the term "plot of raw count profile" or "plot of raw profile" refers to each portion in a region normalized according to the total counts in the region (e.g., Refers to a plot of counts in a genome, portion, chromosome, chromosomal portion of a reference genome, or segment of a chromosome). In some embodiments, profiles may be generated using static windowing, and in certain embodiments, profiles may be generated using sliding windowing.

Profiles generated for a test subject optionally facilitate interpretation of mathematical and/or statistical manipulations of the data set compared to profiles generated for one or more reference subjects, and/ Or present an outcome. In some embodiments, the profile is based on one or more starting assumptions (e.g., maternal nucleic acid contribution (e.g., maternal fraction), fetal nucleic acid contribution (e.g., fetal fraction), reference sample ploidy, etc. or a combination thereof). In certain embodiments, the test profile is often centered around a predetermined value indicative of the absence of a mutation in the gene, and the presence of the gene in the test subject if the test subject carried the mutation in the gene. There are often deviations from given values in areas corresponding to genomic locations where mutations are located. In test subjects at risk for, or suffering from, a medical condition associated with mutation of the gene, the numerical value for the selected portion is significantly altered from the given value for the unaffected genomic location. There is expected. Depending on the starting assumption (e.g., constant or optimized ploidy, constant or optimized fetal fraction, or a combination thereof), a given The threshold or cutoff value or range of thresholds can vary while still providing useful outcomes for determining the presence or absence of genetic mutations. In some embodiments, a profile indicates and/or displays a phenotype.

By way of non-limiting example, the normalized sample and/or reference count profile includes (a) chromosomes, portions, or portions thereof selected from a set of references known not to carry genetic mutations; (b) excluding (e.g., filtering) non-informative portions from raw counts of reference samples; (c) all remaining references The number of reference counts for a genome portion is defined as the total number of remaining counts (e.g., the number of remaining counts after excluding uninformative reference genome portions) for the reference sample, selected chromosome, or selected genomic location. ), thereby generating a normalized reference subject profile, (d) excluding the corresponding portion from the test subject sample, and (e) one or more selected genomes Normalize the remaining test subject counts for the location according to the sum of the median remaining reference counts for the chromosome or chromosomes containing the selected genomic location, thereby yielding a normalized test Generating a profile of interest can be obtained from the raw sequence read data. In certain embodiments, an additional normalization step can be incorporated between (c) and (d) for the whole genome reduced by partial filtering in (b).

A dataset profile can be generated by one or more operations on the counted mapped sequence read data. Some embodiments include: mapping sequence reads and determining (eg, counting) the number of sequence tags that map to each genome portion. A raw count profile is generated from the counted mapped sequence reads. In certain embodiments, the raw count profile from the test subject is for chromosomes, parts, or segments thereof from a set of reference subjects that are known not to carry mutations in the gene. , to present outcomes by comparison with the median reference count profile.

In some embodiments, the sequence read data is optionally filtered to remove noise data or non-informative portions. After filtering, the remaining counts are typically summed to produce a filtered data set. In certain embodiments, a filtered count profile is generated from the filtered data set.

After counting and optionally filtering the sequence read data, the dataset can be normalized to generate levels or profiles. A data set can be normalized by normalizing one or more selected portions according to a suitable normalized reference value. In some embodiments, normalized reference values represent total counts for one or more chromosomes from which portions are selected. In certain embodiments, the normalized reference value is the portion of one or more chromosomes from the reference data set prepared from a set of reference subjects that are known not to carry mutations in the gene. Display one or more corresponding portions where . In some embodiments, the normalized reference value is a portion of one or more chromosomes from a test subject data set prepared from a test subject to be analyzed for the presence or absence of genetic mutations. Display one or more corresponding portions where . In certain embodiments, the normalization process is performed with the help of a static window method, and in some embodiments, the normalization process is performed with the help of a moving window method or a sliding window method. In certain embodiments, profiles containing normalized counts are generated to facilitate classification and/or presentation of outcomes. Outcomes can be presented based on (eg, using such profile plots) plots of profiles that include normalized counts.

Levels In some embodiments, values (eg, numbers, quantitative values) are ascribed to levels. Levels can be determined by any suitable method, operation, or mathematical processing (eg, processed levels). A level is often, or is derived from, a count (eg, normalized count) for a set of parts. In some embodiments, the level of the portion is substantially equal to the total number of counts (eg, counts, normalized counts) that map to the portion. Levels are often determined from counts that have been processed, transformed, or otherwise manipulated by suitable methods, operations, or mathematics known in the art. In some embodiments, the level is derived from the processed counts, non-limiting examples of which are weighted, excluded, filtered, or normalized. adjusted, averaged, derived as an average (eg, average level), added, subtracted, transformed counts, or combinations thereof. In some embodiments, the levels include normalized counts (eg, partial normalized counts). Levels are non-limiting examples of which are normalization for parts, normalization by GC content, median count normalization, linear least-squares regression and non-linear least-squares regression, LOESS (e.g., GC LOESS), LOWESS, It can be a level for counts normalized by a suitable process including PERUN, ChAI, principal component normalization, RM, GCRM, cQn, etc., and/or combinations thereof. Levels can include normalized counts or relative amounts of counts. In some embodiments, the level is for two or more part counts or normalized counts averaged, the level is referred to as the average level. In some embodiments, the level is the level for a set of portions having an average count or average of normalized counts, referred to as the average level. In some embodiments, levels are derived for portions containing raw counts and/or filtered counts. In some embodiments, the level is based on counts, which are raw counts. In some embodiments, the level is associated with an uncertainty value (eg, standard deviation, MAD). In some embodiments, levels are displayed by Z-scores or p-values. As used herein, levels for one or more portions are synonymous with "levels of genome division."

A level for one or more parts is synonymous with "level of genome division" herein. The term "level" is sometimes synonymous with the term "elevation" as used herein. Determination of the meaning of the term "level" can be determined from the context in which it is used. For example, the term "level" often means an increase when used in the context of genomic divisions, profiles, reads, and/or counts. The term "level" often refers to quantity when used in the context of a substance or composition (eg, level of RNA, plexing level). The term "level" often refers to a quantity when used in the context of uncertainty (eg, level of error, confidence, level of deviation, level of uncertainty).

Normalized counts or non-normalized counts for two or more levels (e.g., levels in two or more profiles) are optionally mathematically manipulated according to level (eg, add it, multiply it, average it, normalize it, etc., or combinations thereof). For example, normalized or non-normalized counts for two or more levels can be normalized according to one, some, or all of the levels in the profile. In some embodiments, normalized or non-normalized counts for all levels in the profile are normalized according to one level in the profile. In some embodiments, the normalized or non-normalized counts for the first level in the profile are the normalized or normalized counts for the second level in the profile. Normalize according to the number of counts that are not counted.

Non-limiting examples of levels (e.g., first level, second level) are levels for the set of portions containing the counts processed, the average of the counts, the median, or the portion containing the mean , a level for a set of parts containing normalized counts, etc., or any combination thereof. In some embodiments, the first and second levels in the profile are derived from counts of moieties that map to the same chromosome. In some embodiments, the first and second levels in the profile are derived from counts of moieties that map to different chromosomes.

In some embodiments, levels are determined from normalized or non-normalized counts that map to one or more portions. In some embodiments, the level is determined from normalized or non-normalized counts mapping to two or more parts, where the normalized The count numbers are often nearly the same. The counts (eg, normalized counts) in the set of parts for a level may vary. Within a set of portions for a level, one or more portions may be found that have significantly different counts than within other portions of the set (eg, peaks and/or dips). Any suitable number of normalized or non-normalized counts associated with any suitable number of portions may define a level.

In some embodiments, one or more levels can be determined from normalized or non-normalized counts of all or part of the portion of the genome. Levels can often be determined from all or part of the normalized or non-normalized counts of a chromosome or segment thereof. In some embodiments, two or more counts derived from two or more portions (eg, sets of portions) determine the level. In some embodiments, two or more counts (eg, counts from two or more portions) determine the level. In some embodiments, the number of counts from 2 to about 100,000 fractions determines the level. In some embodiments, 2 to about 50,000, 2 to about 40,000, 2 to about 30,000, 2 to about 20,000, 2 to about 10,000, 2 to about 5000, 2 to about Levels are determined by the number of counts from portions of 2500, 2 to about 1250, 2 to about 1000, 2 to about 500, 2 to about 250, 2 to about 100, or 2 to about 60. In some embodiments, the number of counts from about 10 to about 50 moieties determines the level. In some embodiments, the level is determined by counts from about 20 to about 40 or more moieties. In some embodiments, the levels are about 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21 , 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 45, 50, 55, 60 or more Contains counts from the part. In some embodiments, a level corresponds to a set of parts (eg, a set of parts of a reference genome, a set of chromosome parts, or a set of parts of a chromosomal segment).

In some embodiments, the level is determined for normalized or non-normalized counts of continuous portions. In some embodiments, the contiguous portion (eg, set of portions) represents contiguous segments of a genome or contiguous segments of a chromosome or gene. For example, two or more contiguous portions may represent a sequence assembly of DNA sequences longer than each portion when aligned by joining the portions end-to-end. For example, two or more contiguous portions can represent an intact genome, chromosome, gene, intron, exon, or segment thereof. In some embodiments, levels are determined from a collection (eg, set) of contiguous and/or noncontiguous portions.

Outcomes The methods described herein can result in a determination of the presence or absence of a genetic mutation (e.g., fetal aneuploidy) for a sample, thereby presenting an outcome (e.g., This can present an outcome that determines the presence or absence of genetic mutations (eg, fetal aneuploidy). Genetic mutations may be the gain, loss, and/or alteration (e.g., duplication, acquisition of genetic information (deletion, integration, insertion, mutation, rearrangement, substitution, or methylation aberration) that results in a detectable change in the tested genomic or genetic information relative to the reference , loss, and/or alteration. The presence or absence of genetic mutations can be determined by transforming, analyzing, and/or manipulating sequence reads (e.g., counts, counts of portions of the genome of the reference genome) mapped to portions. can. In some embodiments, determining the outcome comprises analyzing nucleic acid from pregnant females. In certain embodiments, an outcome is counts (e.g., normalized counts, read density, read density profile) obtained from pregnant females, nucleic acids obtained from pregnant females Determined according to the number of counts by

The methods described herein optionally include the presence or absence of fetal aneuploidy (e.g., complete chromosomal aneuploidy, partial chromosomal Determine aneuploidy, or segmental chromosomal abnormalities (eg, mosaics, deletions, and/or insertions). In certain embodiments, the methods described herein detect euploidy or lack of euploidy (non-euploidy) in a sample from a fetus-bearing pregnant female. The methods described herein optionally detect trisomy on one or more chromosomes (eg, chromosome 13, chromosome 18, chromosome 21, or combinations thereof) or segments thereof.

In some embodiments, the presence or absence of a genetic mutation (e.g., fetal aneuploidy) is determined by methods described herein, methods known in the art, or a combination thereof . The presence or absence of a genetic mutation is generally determined from read counts of sequences mapped to portions of the reference genome.

Read densities from a reference are sometimes for nucleic acid samples derived from the same pregnant female from which the test sample is obtained. In certain embodiments, the read density from the reference is for nucleic acid samples derived from one or more pregnant females different from the female from which the test sample was obtained. In some embodiments, comparing the read density and/or read density profile from the first set of portions from the test subject to the read density and/or read density profile from the second set of portions; Here the second set of parts is different from the first set of parts. In some embodiments, comparing the read density and/or read density profile from the first set of portions from the test subject to the read density and/or read density profile from the second set of portions; Here the second set of parts is either from the test subject or from a reference subject that is not the test subject. In a non-limiting example, where the first set of portions is in chromosome 21 or a segment thereof, the second set of portions is in another chromosome (e.g., chromosome 1, chromosome 13, chromosome 14 Chromosome, Chromosome 18, Chromosome 19, a segment thereof, or a combination of the foregoing). References are often located in chromosomes or segments thereof that are typically euploid. For example, chromosomes 1 and 19 are euploid due to the high rate of early fetal mortality associated with chromosome 1 and chromosome 19 aneuploidies in the fetus. often the body. A measure of uncertainty between the read density and/or read density profile from the test subject and the reference can be generated and/or compared. The presence or absence of a genetic mutation (eg, fetal aneuploidy) is sometimes determined without comparing read densities and/or read density profiles from test subjects to a reference.

In certain embodiments, the reference comprises read densities and/or read profiles for the same set of parts as for the test subject, where the read densities for the reference are one or more reference samples (e.g., often from multiple reference samples derived from multiple reference subjects). Reference samples are often derived from one or more pregnant females different from the female from which the test sample is obtained.

A measure of uncertainty for the read density and/or read profile of the test subject and/or reference can be generated. In some embodiments, a measure of uncertainty is determined for the read density and/or read profile of the test subject. In some embodiments, a measure of uncertainty is determined for a reference read density and/or read profile. In some embodiments, the uncertainty measure is determined from the entire read density profile or a subset of portions within the read density profile.

In some embodiments, the reference sample is euploid for the selected segment of the genome, and a measure of uncertainty between the test profile and the reference profile is evaluated for the selected segment. In some embodiments, the determination of the presence or absence of a genetic mutation is based on the deviation (e.g., a measure of deviation, MAD) number. In some embodiments, the number of deviations between the test profile and the reference profile is greater than about 1, greater than about 1.5, greater than about 2, greater than about 2.5, greater than about 2.6, greater than about 2.7 , greater than about 2.8, greater than about 2.9, greater than about 3, greater than about 3.1, greater than about 3.2, greater than about 3.3, greater than about 3.4, greater than about 3.5, greater than about 4 , greater than about 5, or greater than about 6 determines the presence of a mutation in the gene. For example, optionally, the presence of a mutation in a gene is determined if the test profile and the reference profile differ by more than 3 on a measure of deviation (eg, 3 sigma, 3 MAD). In some embodiments, fetal chromosomal aneuploidy (e.g., fetal chromosomal trisomy) is determined. A deviation of more than 3 between the test profile and the reference profile is often indicative of a non-euploid test subject (eg, the presence of genetic mutations) for the selected segment of the genome. Optionally, a test profile that significantly exceeds a reference profile for a selected segment of the genome, where the reference is euploid for the selected segment, determines trisomy. In some embodiments, fetal chromosomal aneuploidy if the read density profile obtained from the pregnant female is less than 3 on a deviation measure (e.g., 3 sigma, 3 MAD) from the reference profile for the selected segment. The presence of sex (eg, fetal monosomy) is determined. In some cases, monosomy is determined by a test profile significantly below a reference profile indicative of euploidy.

In some embodiments, the number of deviations between the test profile and the reference profile for the selected segment of the genome is less than about 3.5, less than about 3.4, less than about 3.3, less than about 3.2 , less than about 3.1, less than about 3.0, less than about 2.9, less than about 2.8, less than about 2.7, less than about 2.6, less than about 2.5, less than about 2.0, about Absence of mutation of the gene is determined if less than 1.5, or less than about 1.0. For example, optionally, the absence of mutation in a gene is determined if the test profile differs from the reference profile by less than 3 on a measure of deviation (eg, 3 sigma, 3 MAD). In some embodiments, the absence of fetal chromosomal aneuploidy (e.g., , fetal euploid) is determined. In some embodiments, (e.g., a deviation of less than 3 (e.g., 3 sigma in standard deviation) between the test profile and the reference profile is euploid segment of the genome (e.g., absence of mutation in gene A measure of deviation between a test profile for a test sample and a reference profile for one or more reference subjects can be plotted and visualized (e.g., a z-score plot).

Any other suitable reference may be used to determine the presence or absence of a genetic mutation (or euploid or non-euploid determination) for a test region of a test sample (e.g., a segment of the genome being tested). ) can be factored with the test profile for In some embodiments, the fetal fraction determination can be factored by sequence read counts (eg, read density) to determine the presence or absence of genetic mutations. For example, read densities and/or read density profiles can be normalized according to fetal fraction before comparing and/or determining outcomes. Suitable processes for quantifying the fetal fraction can be utilized, non-limiting examples of which include mass spectrometric processes, sequencing processes, or combinations thereof.

In some embodiments, the determination of the presence or absence of a genetic mutation (eg, fetal aneuploidy) is determined according to decision zones. In certain embodiments, a determination ( For example, making a decision (eg, outcome) that determines the presence or absence of mutations in a gene. In some embodiments, decision zones are defined according to a collection of values (eg, read density profile and/or uncertainty measure) obtained from the same patient sample. In certain embodiments, the decision regions are defined according to a collection of values derived from the same chromosome or segment thereof. In some embodiments, the decision area based on the genetic mutation determination is defined according to a measure of uncertainty (eg, high confidence level, eg, low uncertainty measure) and/or fetal fraction.

In some embodiments, the decision range is determined by determination of genetic mutation and about 2.0% or more, about 2.5% or more, about 3% or more, about 3.25% or more. , about 3.5% or more, about 3.75% or more, or about 4.0% or more of the fetal fraction. For example, in some embodiments, the test sample from which the test profile was derived was a fetal fraction of 2% or greater or 4% or greater for test samples obtained from fetal-bearing pregnant females. a determination that the fetus contains trisomy 21 based on the comparison of the test profile and the reference profile. For example, in some embodiments, the test sample from which the test profile was derived was a fetal fraction of 2% or greater or 4% or greater for test samples obtained from fetal-bearing pregnant females. a determination that the fetus is euploid based on the comparison of the test profile and the reference profile. In some embodiments, the decision range is about 99% or more, about 99.1% or more, about 99.2% or more, about 99.3% or more, about 99.4% or more, about 99.5% or more, about 99.6% or more, about 99.7% or more, about 99.8% or more, or about 99.9% or more Defined by the reliability level. In some embodiments, decisions are made without using decision zones. In some embodiments, decision zones and additional data or information are used to make decisions. In some embodiments, decisions are made based on comparisons without the use of decision zones. In some embodiments, the determination is made based on viewing the profile (eg, viewing lead density).

In some embodiments, there is no decision zone if no decision is made. In some embodiments, the no-decision zone is defined by a value or collection of values that indicate low accuracy, high risk, high error, low confidence level, high uncertainty measure, etc., or a combination thereof. . In some embodiments, the no-decision zone is about 5% or less, about 4% or less, about 3% or less, about 2.5% or less, about 2.0% or less , in part defined by a fetal fraction of about 1.5% or less, or about 1.0% or less.

Gene mutations are sometimes associated with medical conditions. An outcome that determines a genetic mutation is optionally an outcome that determines the presence or absence of a condition (e.g., a medical condition), disease, syndrome, or abnormality, or a condition, disease, syndrome, or abnormality ( For example, non-limiting examples listed in Table 1). In certain embodiments, diagnosing comprises evaluating an outcome. Outcomes that determine the presence or absence of a condition (e.g., medical condition), disease, syndrome, or abnormality by the methods described herein are optionally further examined (e.g., karyotyping and/or or by amniocentesis) can be independently verified. Analysis and processing of data may present one or more outcomes. As used herein, the term "outcome" when referring to the result of data processing that facilitates determining the presence or absence of genetic mutations (e.g., aneuploidy, copy number variation) There is In certain embodiments, the term "outcome" as used herein refers to a conclusion that predicts and/or determines the presence or absence of genetic variation (e.g., aneuploidy, copy number variation). Point. In certain embodiments, the term "outcome" as used herein refers to the risk of the presence or absence of a genetic mutation (e.g., aneuploidy, copy number variation) in a subject (e.g., fetus). Refers to a conclusion that predicts and/or determines sex or probability. Diagnosis optionally includes the use of outcomes. For example, a medical practitioner can analyze the outcome and present a diagnosis based on the outcome, or based in part on the outcome. In some embodiments, determining, detecting, or diagnosing a condition, syndrome, or abnormality (e.g., listed in Table 1) includes using outcomes that determine the presence or absence of genetic mutations. In some embodiments, the counted mapped sequence reads or their conversion-based outcomes determine the presence or absence of genetic mutations. In certain embodiments, outcomes generated utilizing one or more methods (e.g., data processing methods) described herein are one or more conditions listed in Table 1; Determine the presence or absence of a syndrome, or anomaly. In certain embodiments, diagnosis includes determining the presence or absence of a condition, syndrome, or abnormality. Diagnosis often involves determining genetic mutations as the nature and/or cause of a condition, syndrome, or disorder. In certain embodiments the outcome is not a diagnosis. In the context of considering one or more probabilities, outcomes often include one or more numerical values generated using the processing methods described herein. Risk or probability considerations include measures of uncertainty, confidence level, sensitivity, specificity, standard deviation, coefficient of variation (CV) and/or confidence level, Z-score, chi value, phi value, ploidy value. , matched fetal fraction, area ratio, median level, etc., or combinations thereof. Probability studies can facilitate determining whether a subject is at risk of having a genetic mutation or whether a subject has a genetic mutation, determining the presence or absence of a genetic disorder. Outcomes to be evaluated often include such considerations.

Outcomes are sometimes phenotypes. Outcomes are optionally associated with a relevant confidence level (e.g., a measure of uncertainty, e.g., fetus is positive for trisomy 21 at a 99% confidence level, test subject is positive for trisomy 21 at a 95% confidence level, negative for cancer associated with genetic mutations). Different methods of generating outcome values can lead to different kinds of results in some cases. Generally, there are four types of possible scores or judgments that can be made based on outcome values generated using the methods described herein: true positives, false positives, true negatives, and false negatives. As used herein, the terms "score", "scores", "call" and "calls" refer to the effect that a particular gene mutation has on a subject/sample. Refers to calculating the probability of existence or non-existence. Score values can be used, for example, to determine mutations, differences, or ratios of mapped sequence reads that may correspond to mutations in the gene. For example, by calculating positive scores for selected gene mutations or portions from the dataset relative to the reference genome, gene mutations (e.g., cancer , preeclampsia, trisomy, monosomy, etc.). In some embodiments, outcomes include read densities, read density profiles, and/or plots (eg, plots of profiles). In embodiments where an outcome includes a profile, any suitable profile or combination of profiles can be used for the outcome. Non-limiting examples of profiles that may be used for outcomes include z-score profiles, p-value profiles, chi-value profiles, phi-value profiles, etc., and combinations thereof.

Outcomes generated to determine the presence or absence of a gene mutation may, in some cases, include null results (e.g., data points between two clusters, values for both the presence and absence of a gene mutation). data sets with plots of profiles that are not similar to plots of profiles for subjects with or without mutations in the sought gene). In some embodiments, an outcome that is indicative of a null result is also a decision-making result, the decision being repeated and/or generating additional information and/or data to determine the presence or absence of mutations in the gene. May include need for analysis.

In some embodiments, outcomes can be generated after performing one or more processing steps described herein. In certain embodiments, an outcome is generated as a result of one of the processing steps described herein, and in some embodiments an outcome is each statistical manipulation of the data set and/or It can be generated after performing each mathematical operation. Outcomes for determining the presence or absence of a genetic variant may include, without limitation, probability (e.g., odds ratio, p-value), likelihood, in-cluster or out-of-cluster value, above threshold value or above threshold value. Suitable forms, including a value below, a value within a range (e.g., a threshold range), a measure of variance or a value with confidence, or a risk factor associated with the presence or absence of a genetic mutation for a subject or sample. can be expressed as In certain embodiments, comparisons between samples allow confirmation of sample identity (e.g., repeated samples and/or mixed samples (e.g., mislabeled samples, combined samples, etc.). ) can be identified).

In some embodiments, the outcome includes a value above or below a predetermined threshold or cutoff value and/or a level of uncertainty or confidence level associated with the value. In certain embodiments, the predetermined threshold or cutoff value is an expected level or range of expected levels. Outcomes can also describe the assumptions used in data processing. In certain embodiments, the outcome is a value that is within or outside a predetermined range of values (e.g., a threshold range) and the associated level of uncertainty or Includes confidence level. In some embodiments, the outcome is equal to a predetermined value (e.g., equal to 1, equal to zero) or a value within a range of predetermined values, and is equal to or within a range of values or its associated level of uncertainty or confidence level for that value that is out of range. Outcomes are sometimes represented graphically as plots (eg, profile plots).

As noted above, outcomes can be characterized as true positives, true negatives, false positives, or false negatives. As used herein, the term "true positive" refers to a subject who has been correctly diagnosed as having a genetic mutation. As used herein, the term "false positive" refers to a subject that is falsely identified as having a genetic mutation. As used herein, the term "true negative" refers to a subject that has been correctly identified as having no genetic mutation. As used herein, the term "false negative" refers to a subject that is falsely identified as having no genetic mutation. Two measures of efficacy for any given method are: (i) a sensitivity value, which is generally the percentage of positives expected that are correctly identified as positives; It can be calculated based on the incidence rate of specificity values, which is the percentage of positive negatives identified correctly as negatives.

In certain embodiments, one or more of sensitivity, specificity, and/or confidence levels are expressed as percentages. In some embodiments, the percentages are greater than about 90% (e.g., about 90, 91, 92, 93, 94, 95, 96, 97, 98, or 99%, or greater than 99% for each variable independently). (eg, about 99.5% or more, about 99.9% or more, about 99.95% or more, about 99.99% or more)). In some embodiments, the coefficient of variation (CV) is expressed as a percentage, and in some cases the percentage is about 10% or less (e.g., about 10, 9, 8, 7, 6, 5, 4, 3 , 2, or 1%, or less than 1% (e.g., about 0.5% or less, about 0.1% or less, about 0.05% or less, about 0.01% or less) ). In certain embodiments, probabilities (eg, the probability that a particular outcome is not due to chance) are expressed as Z-scores, p-values, or t-test results. In some embodiments, the measured variance, confidence interval, sensitivity, specificity, etc. (e.g., collectively referred to as reliability parameters) for the outcome are one or more can be generated using data processing operations of A specific example of generating outcomes and associated confidence levels is found in International Patent Application No. PCT /US12/59123 (WO2013/052913).

As used herein, the term "sensitivity" refers to the number of true positives divided by the number of true positives plus the number of false negatives, where sensitivity ( sens) can be in the range 0≦sens≦1. As used herein, the term "specificity" refers to the number of true negatives divided by the number of true negatives plus the number of false positives, where sensitivity (spec) can be in the range 0≦spec≦1. In some embodiments, methods are optionally selected in which the sensitivity and specificity are equal to or near 1 or 100% (eg, between about 90% and about 99%). In some embodiments, methods with sensitivity equal to 1 or 100% are selected, and in certain embodiments, sensitivity is near 1 (e.g., about 90% sensitivity, about 91% sensitivity, about 92% sensitivity, about 93% sensitivity, about 94% sensitivity, about 95% sensitivity, about 96% sensitivity, about 97% sensitivity, about 98% sensitivity, or about 99% sensitivity) choose a method. In some embodiments, methods are selected that have a specificity equal to 1 or 100%, and in certain embodiments, a specificity near 1 (e.g., about 90% specificity, about 91% specificity, about 92% specificity, about 93% specificity, about 94% specificity, about 95% specificity, about 96% specificity, about 97% specificity, about 98% specificity or a specificity of about 99%).

In some embodiments, the presence or absence of genetic mutation (eg, chromosomal aneuploidy) is determined in the fetus. In such embodiments, the presence or absence of a fetal genetic mutation (eg, fetal chromosomal aneuploidy) is determined.

In certain embodiments, the presence or absence of genetic mutations (eg, chromosomal aneuploidy) is determined for a sample. In such embodiments, the presence or absence of genetic mutations in a sample nucleic acid (eg, chromosomal aneuploidy) is determined. In some embodiments, a detected or undetected mutation is present in sample nucleic acid from one source but not in sample nucleic acid from another source. Non-limiting examples of sources include placental nucleic acid, fetal nucleic acid, maternal nucleic acid, cancer cell nucleic acid, non-cancer cell nucleic acid, etc., and combinations thereof. In a non-limiting example, a detected or undetected mutation of a particular gene is (i) present in placental nucleic acid, but not in fetal nucleic acid, nor in maternal nucleic acid; (ii) present in fetal nucleic acid but absent in maternal nucleic acid; or (iii) present in maternal nucleic acid but absent in fetal nucleic acid.

The presence or absence of genetic mutations and/or associated medical conditions (eg, outcomes) are often provided by outcome modules. The presence or absence of a genetic mutation (eg, aneuploidy, fetal aneuploidy, copy number variation) is identified, in some embodiments, by an outcome module or a machine comprising an outcome module. Outcome modules may specialize in determining specific genetic mutations (eg, trisomy, trisomy 21, trisomy 18). For example, an outcome module that identifies trisomy 21 may be different and/or different than an outcome module that identifies trisomy 18. In some embodiments, an outcome module or machine comprising an outcome module is required to identify mutations in genes or definitive outcomes of mutations in genes (eg, aneuploidy, copy number variation). In certain embodiments, the outcomes are transferred from the outcomes module to the display module, where the outcomes are provided by the display module.

A genetic mutation or conclusive outcome of a genetic mutation identified by the methods described herein is independently verified by further testing (e.g., by targeted sequencing of maternal and/or fetal nucleic acids). obtain. Outcomes are typically provided to health care professionals (eg, laboratory technicians or administrators; physicians or assistants). In certain embodiments, the outcomes are provided on a suitable visual medium (eg, a machine peripheral or component, eg, a printer or display). In some embodiments, the outcome determining the presence or absence of a genetic mutation is presented to a healthcare professional in the form of a report, and in certain embodiments, the report includes the outcome value and the associated including the presentation of reliability parameters for In general, outcomes can be presented in a suitable format that facilitates determination of the presence or absence of genetic mutations and/or medical conditions. Non-limiting examples of formats suitable for use in reporting and/or presenting datasets or reporting outcomes include digital data, graphs, 2D graphs, 3D graphs, and 4D graphs, photographs (e.g., jpg , bitmap (e.g. bmp), pdf, tiff, gif, raw, png, etc., or any suitable format), statistical charts, charts, tables, bar charts, pie charts, schematics, flowcharts, scatterplots, maps, histograms, density Including diagrams, function graphs, circuit diagrams, block diagrams, bubble maps, signal space diagrams, contour diagrams, cartograms, radar charts, Venn diagrams, nomograms, etc., and combinations of the foregoing. Various examples of outcome displays are illustrated in the figures and described in the examples.

In certain embodiments, the generation of an outcome can be considered the conversion of a subject's intracellular nucleic acid representations, such as nucleic acid sequence read data. For example, analyzing nucleic acid sequence reads from a subject to generate chromosomal profiles and/or outcomes can be considered the conversion of relatively small sequence read fragments into relatively large chromosomal structural representations. can be done. In some embodiments, the outcome is the pre-existing structure (e.g., genomic , chromosomes or segments thereof) resulting from conversion into representations. In some embodiments, the outcome is the conversion of sequence reads from a first subject (e.g., a pregnant female) into a composite representation of a structure (e.g., genome, chromosome or segment thereof), and Including a second transformation of the composite representation that provides a representation of structures present within one subject (eg, a pregnant female) and/or within a second subject (eg, a fetus).

Sex Chromosome Outcomes In some embodiments, the outcome relates to mutations in sex chromosome genes. Sex chromosome gene mutations are described, for example, in International Patent Application Publication No. WO2013/192562, the entire contents of which are incorporated herein by reference, including all documents, tables, formulas, and drawings. ing. In some embodiments, the outcome is determination of sex chromosome karyotype, detection of sex chromosome aneuploidy, and/or determination of fetal sex. Some sex chromosome aneuploidy (SCA) conditions include Turner syndrome [45,X], trisomy X [47,XXX], Klinefelter syndrome [47,XXY], and [47,XYY] syndrome (sometimes Jacob's syndrome), including but not limited to:

Assessment of sex chromosome variation is, in some embodiments, based on segregation of transforming sequence read counts for chromosome X and chromosome Y. Transformation of sequence read counts can include, for example, chromosome X representation and chromosome Y representation, and/or Z-scores based on such representations. Nucleotide sequence read count conversion (e.g., PERUN normalized A two-dimensional plot of read counts or z-scores based on principal component normalized read counts) produces a planar field of plotted points, each of which can be carved into a region specific to a particular karyotype. For example, determining the sex chromosome karyotype for a given sample can be accomplished by determining in which region of the planar field the plot points for that sample lie.

Certain methods described herein can be useful for generating plots with well-defined regions (e.g., with sharp boundaries, high resolution) for specific karyotypic variations. Methods that can help generate high-resolution plots include normalization of sequence read counts, selection of informative portions (i.e. bins) for chromosome X and chromosome Y, non-reportable (i.e. , “no-decision” regions) and additional normalization of chromosome X and chromosome Y levels. Normalization of sequence reads and further normalization of levels are described herein, e.g., sequence reads mapped to chromosome X and/or chromosome Y, and/or chromosome X and/or It may include PERUN normalization and/or principal component normalization of levels (eg, chromosome representations) for Y. The selection of informative moieties for chromosome X and chromosome Y is described, for example, in International Patent Application Publication No. WO2013/192562 and includes filtering parameters such as cross-validation parameters, mappability, reproducibility, and/or evaluation such as male versus female segregation.

Use of Outcomes A healthcare professional or other qualified person who receives a report containing outcomes that determine the presence or absence of mutations in one or more genes may use the data presented in the report to: A determination can be made about the condition of the test subject or patient. In some embodiments, medical professionals can make recommendations based on presented outcomes. In some embodiments, a healthcare professional or qualified individual determines the presence of a genetic mutation in a test subject or patient based on one or more outcome values and associated reliability parameters presented in the report. Or a judgment or score for non-existence can be presented. In certain embodiments, visual observation of the presented report is used to manually generate scores or make judgments by medical personnel or qualified personnel. In certain embodiments, an automated routine, optionally embedded within the software, generates a score or makes a judgment, prior to providing the information to a test subject or patient, by a healthcare professional or A qualified person reviews the accuracy. The term "receiving a report" as used herein means that, upon review, a medical professional or other qualified person determines the presence or absence of a genetic mutation in a test subject or patient. It refers to obtaining through communication means, textual representations, and/or graphical representations of outcomes, including those that allow The report can be computer generated, can be generated by manual data entry, and can be submitted electronically (e.g., via the Internet, via computer, via fax, at the same physical facility or at a different physical facility). from one network location at a facility to another) or by other methods of sending or receiving data (eg, mail, courier, etc.). In some embodiments, outcomes are transmitted to healthcare professionals by any suitable medium including, but not limited to, verbal, written, or file form. A file can be, for example, without limitation, an audio file, a computer readable file, a paper file, a laboratory file, or a medical record file.

As used herein, the term "presenting an outcome" and its grammatical equivalents also refer to methods for obtaining such information, including, without limitation, examining information. It may also refer to a method that includes obtaining from a laboratory (eg, laboratory file). A laboratory file can be created by a laboratory that has performed one or more assays and can be created by one or more data processing steps that determine the presence or absence of a medical condition. The laboratory may be co-located with the medical personnel identifying the presence or absence of the medical condition from laboratory files, or it may be in a different location (eg, another country). For example, a laboratory file can be created at one location and transmitted to another location where the information therein is transmitted to the pregnant female subject. In certain embodiments, the laboratory file may be in physical form or in electronic form (eg, computer readable form).

In some embodiments, the outcome can be presented from the laboratory to a healthcare professional, physician, or qualified person, who can make a diagnosis based on the outcome. can be lowered. In some embodiments, outcomes can be presented from the laboratory to a healthcare professional, physician, or qualified person, who can provide additional data and/or information. , as well as other outcomes, can be based in part on the outcome to make a diagnosis.

A healthcare professional or qualified person can make appropriate recommendations based on one or more of the outcomes presented in the report. Non-limiting examples of recommendations that may be made based on the outcome report provided include surgery, radiation therapy, chemotherapy, genetic counseling, postnatal treatment solutions (e.g., life planning, long-term care, pharmaceuticals, symptomatic treatment), abortion, organ transplantation, blood transfusion, etc., or combinations of the foregoing. In some embodiments, the recommendations are based on the outcome-based classification presented (e.g., Down syndrome, Turner syndrome, medical conditions associated with genetic mutations at T13, medical conditions associated with genetic mutations at T18) depends on

A laboratory official (e.g., laboratory manager) determines the value (e.g., test profile, reference profile, level of deviation) can be analyzed. For determinations of the presence or absence of genetic mutations that are subtle or questionable, laboratory personnel may reorder the same test and/or Different tests (eg, karyotyping and/or amniocentesis in the case of determination of fetal aneuploidy) can also be ordered using the same or different sample nucleic acids from which they originate.

Genetic Mutations and Medical Conditions The presence or absence of genetic differences can be determined using the methods, machines or devices described herein. In certain embodiments, the presence or absence of mutations in one or more genes is determined by results provided by the methods, machines and devices described herein. A genetic variation is generally a specific genetic phenotype present in a particular individual, and often a genetic variation is present in a statistically significant subpopulation of individuals. . In some embodiments, the genetic mutation is a chromosomal abnormality (e.g., aneuploidy, duplication of one or more chromosomes, loss of one or more chromosomes), partial chromosomal abnormality or mosaicism (e.g., loss or gain of one or more segments of a chromosome), translocations, and inversions, each of which is described in more detail herein. Non-limiting examples of genetic mutations include one or more deletions (e.g., microdeletion), duplications (e.g., microduplications), insertions, mutations, polymorphisms (e.g., single nucleotide polymorphisms), Fusions, repeats (eg, short tandem repeats), different methylation sites, different methylation patterns, etc., and combinations thereof. Insertions, repeats, deletions, duplications, mutations, or polymorphisms can be of any length, and in some embodiments from about 1 base or base pair (bp) to about 250 megabases in length. (Mb). In some embodiments, an insertion, repeat, deletion, duplication, mutation, or polymorphism is from about 1 base or base pair (bp) to about 50,000 kilobases (kb) in length (e.g., about 10 bp, 50 bp, 100 bp, 500 bp, 1 kb, 5 kb, 10 kb, 50 kb, 100 kb, 500 kb, 1000 kb, 5000 kb, or 10,000 kb in length).

Gene mutations can also be deletions. In certain embodiments, the deletion is a mutation in which part of the chromosome or DNA sequence is missing (eg, genetic abnormality). A deletion is often a loss of genetic material. Any number of nucleotides can be deleted. Deletions can include deletions of one or more entire chromosomes, segments of chromosomes, alleles, genes, introns, exons, any noncoding region, any coding region, segments thereof, or combinations thereof. Deletions can include microdeletion. Deletions can include deletions of single bases.

Gene mutation may also be gene duplication. In certain embodiments, duplications are mutations (eg, genetic aberrations) in which a portion of a chromosomal or DNA sequence is copied and reinserted into the genome. In certain embodiments, a gene duplication (eg, duplication) is any duplication of a DNA region. In some embodiments, the duplication is a nucleic acid sequence, often tandemly repeated, within a genome or chromosome. In some embodiments, the duplication is a copy of one or more entire chromosomes, segments of chromosomes, alleles, genes, introns, exons, any noncoding region, any coding region, segments thereof, or combinations thereof. can include Overlaps can include microduplications. A duplication may include one or more duplicate copies of a nucleic acid. A duplication may be characterized as a gene region that has been repeated one or more times (eg, 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 times). Duplications can range from small regions (a few thousand base pairs) to in some cases entire chromosomes. Duplications frequently occur as a result of errors in homologous recombination or due to retrotransposon events. The duplication has been associated with certain species of proliferative disease. Duplications can be characterized using genomic microarrays or comparative genetic hybridization (CGH).

Gene mutations may also be insertions. Insertions may be additions of one or more nucleotide base pairs to a nucleic acid sequence. The insertion may also be microinsertion. In certain embodiments, an insertion comprises the addition of a segment of a chromosome to the genome, chromosome, or segment thereof. In certain embodiments, insertions include additions of alleles, genes, introns, exons, any non-coding regions, any coding regions, segments or combinations thereof to the genome or segments thereof. In certain embodiments, an insertion comprises an addition (eg, an insertion) of a nucleic acid of unknown origin to a genome, chromosome, or segment thereof. In certain embodiments, an insertion comprises a single base addition (eg, insertion).

As used herein, a "copy number variation" is generally a class or type of genetic variation or chromosomal abnormality. Copy number variations can be deletions (eg, microdeletions), duplications (eg, microduplications), or insertions (eg, microinsertions). Often, the prefix "microscopic" as sometimes used herein is a segment of nucleic acid less than 5 Mb in length. Copy number variations can involve one or more deletions (eg, microdeletions), duplications, and/or insertions (eg, microduplications, microinsertions) of segments of the chromosome. In certain embodiments, a duplication includes an insertion. In certain embodiments the insertion is a duplication. In certain embodiments, an insertion is not a duplication.

In some embodiments, the copy number variation is a fetal copy number variation. In many cases, the fetal copy number variation is a copy number variation within the genome of the fetus. In some embodiments, the copy number variation is a maternal and/or fetal copy number variation. In certain embodiments, the maternal and/or fetal copy number alterations are associated with pregnant females (e.g., fetal-bearing female subjects), parity-bearing female subjects, or fetal-bearing female subjects. is a copy number variation within the genome of A copy number variation can be a heterozygous copy number variation, where the variation (eg, duplication or deletion) is present on one allele of the genome. A copy number variation may be a homozygous copy number variation, where the variation is present in both alleles of the genome. In some embodiments, the copy number variation is a heterozygous or homozygous fetal copy number variation. In some embodiments, the copy number variation is a heterozygous or homozygous maternal and/or fetal copy number variation. The copy number variation may be present in the maternal and fetal genome, present in the maternal genome but not in the fetal genome, or present in the fetal genome but not in the maternal genome.

"Ploidy" refers to the number of chromosomes present in a fetus or mother. In certain embodiments, "polyploidy" is the same as "chromosomal polyploidy." In humans, for example, autosomes often exist in pairs. For example, in the absence of genetic mutations, most humans have two of each autosome (eg, chromosomes 1-22). In humans there is a normal complement for two autosomes, which is often called euploid or diploid. "Microploidy" is similar in meaning to polyploidy. "Microploidy" often refers to the polyploidy of a segment of a chromosome. The term "microploidy" refers to the presence or absence of intrachromosomal copy number variations (e.g., deletions, duplications, and/or insertions) (e.g., homozygous or heterozygous deletions, duplications, or insertion, etc. or its absence).

In certain embodiments, the microploidy of the fetus matches the microploidy of the fetal mother (eg, a pregnant female subject). In certain embodiments, the microploidy of the fetus matches the maternal microploidy of the fetus, and both the mother and the fetus have the same heterozygous copy number mutation, homozygous copy number mutation or both are euploid. In certain embodiments, the microploidy of the fetus is different from the microploidy of the fetus' mother. For example, a fetal microploidy is heterozygous for a copy number mutation, a mother is homozygous for a copy number mutation, and a fetal microploidy is a In some cases, microploidy is not consistent with (eg, not equal to).

A genetic mutation identified to be present or absent in a subject is associated with a medical condition in certain embodiments. Accordingly, the techniques described herein can be used to identify the presence or absence of mutations in one or more genes associated with a medical condition or disease state. Non-limiting examples of medical conditions include intellectual disability (e.g., Down's syndrome), cell proliferative disorders (e.g., cancer), presence of microbial nucleic acids (e.g., viruses, bacteria, fungi, yeast), and pre-eclampsia. disease-related conditions.

Non-limiting examples of genetic mutations, medical conditions and medical conditions are described below.

Fetal Sex In some embodiments, prediction of fetal sex or a sex-related disorder (e.g., sex chromosome aneuploidy) can be determined by the methods, machines and/or devices described herein. . Gender determination is generally based on sex chromosomes. In humans, there are two sex chromosomes, the X and Y chromosomes. The Y chromosome contains the gene, SRY, that triggers the embryo to develop as a male. The Y chromosome of humans and other mammals also contains other genes required for normal sperm production. Individuals with XX are female and XY are male, non-limiting mutations often called sex chromosome aneuploidies include X0, XYY, XXX, and XXY. In certain embodiments, males have two X chromosomes and one Y chromosome (XXY; Klinefelter syndrome), or one X chromosome and two Y chromosomes (XYY syndrome; Jacobs syndrome), and one Part females have either three X chromosomes (XXX; triple X syndrome) or a single X chromosome instead of two (X0; Turner syndrome). In certain embodiments, only some cells within an individual are affected by sex chromosome aneuploidy, sometimes referred to as mosaicism (eg, Turner's mosaicism). Other cases include those in which SRY is damaged (resulting in XY females) or copied to X (resulting in XX males).

In certain cases, it may be beneficial to determine the sex of the fetus in utero. For example, a patient (e.g., a pregnant female) with a family history of one or more sex-related disorders determines the sex of a pregnant fetus to help assess the risk of a fetus inheriting such disorders. Sometimes you want to. Sex-related disorders include, but are not limited to, X-linked and Y-linked disorders. X-linked disorders include X-linked recessive disorders and X-linked dominant disorders. Examples of X-linked recessive disorders include, but are not limited to, immune disorders such as chronic granulomatous disease (CYBB), Wiskott-Aldrich syndrome, X-linked severe combined immunodeficiency, X-linked agammaglobulinemia, 1 type hyper-IgM syndrome, IPEX, X-linked lymphoproliferative disorders, properdin deficiency), blood disorders (e.g. hemophilia A, hemophilia B, X-linked sideroerythroblastic anemia), endocrine disorders (e.g. , androgen insensitivity syndrome/Kennedy disease, KAL1 Kallman syndrome, X-linked congenital adrenal hypoplasia), metabolic disorders (eg, ornithine transcarbamylase deficiency, oculocerebral renal syndrome, adrenoleukodystrophy, glucose-6-phosphate) dehydrogenase deficiency, pyruvate dehydrogenase deficiency, Danon disease/type IIb glycogen storage disease, Fabry disease, Hunter syndrome, Lesch-Nyhan syndrome, Menkes disease/occipital Horn syndrome), nervous system disorders (e.g. Coffin-Lowry syndrome, MASA syndrome, X-linked alpha-thalassemia mental retardation syndrome, Siderius X-linked mental retardation syndrome, color blindness, leukoplakia, Norrier disease, Colloideremia, Charcot-Marie-Tooth disease (CMTX2-3), Peliszeus-Merzbacher disease, SMAX2), skin and related tissue disorders (e.g. dyskeratosis congenita, anhidrotic ectodermal dysplasia (EDA), X-linked ichthyosis, X-linked corneal endothelial degeneration), neuromuscular disorders (e.g. Becker muscular dystrophy/Duchenne muscular dystrophy, central nuclear myopathy (MTM1), Conrady-Hunelmann syndrome, Emery-Dreyfus muscular dystrophy 1), urinary system disorders (e.g. Alport syndrome, Dent's disease, X-linked diabetes insipidus), bone/dental disorders ( Amelax hypoplasia), and other disorders (eg, Barth Syndrome, McLeod Syndrome, Smith-Feynman-Myers Syndrome, Simpson-Gorabi-Behmer Syndrome, Mohr-Tranebjaerg Syndrome, Nasal Finger Hearing Syndrome). Examples of X-linked dominant disorders include, but are not limited to, X-linked hypophosphatemia, focal cutaneous hypoplasia, fragile X syndrome, Aicardi syndrome, incontinence pigment, Rett syndrome, CHILD syndrome, Lujan-Fryns syndrome, and oral-facial-finger-toe syndrome 1. Examples of Y-linked disorders include, without limitation, male infertility, retinitis pigmentosa, and azoospermia.

Chromosomal Abnormalities In some embodiments, the presence or absence of fetal chromosomal abnormalities can be determined using the methods, machines and/or devices described herein. Chromosomal abnormalities include, but are not limited to, the gain or loss of entire chromosomes or regions of chromosomes containing one or more genes. Chromosomal abnormalities include one or more nucleotide sequences (e.g., one or more genes) deletions and/or duplications. The terms "chromosomal abnormality" or "chromosomal aneuploidy" as used herein refer to the divergence between the chromosomal structure of interest and the normal homologous chromosomal structure. The term "normal" refers to the predominant karyotype or banding pattern found in healthy individuals of a particular species, e.g. Point. Different organisms also vary widely in their chromosomal complements, and the term "chromosomal aneuploidy" does not refer to a specific number of chromosomes, but rather abnormal chromosomal content within one or more of a given cell of an organism. refers to a situation where In some embodiments, the term "chromosomal aneuploidy" as used herein refers to an imbalance of genetic material caused by the loss or gain of all or part of a chromosome. A "chromosomal aneuploidy" can refer to one or more deletions and/or insertions of segments of a chromosome. The term "euploid" refers, in some embodiments, to the normal complement of chromosomes.

The term "monosomy" as used herein refers to the absence of one chromosome of the normal complement. Partial monosomy can occur in unbalanced translocations or deletions in which only segments of the chromosome are present in a single copy. Sex chromosome monosomy (45, X) causes, for example, Turner's syndrome. The term "disomy" refers to the presence of two copies of a chromosome. For organisms such as humans that have two copies of each chromosome (diploid or "euploid" organisms), disomy is a normal condition. For organisms that normally have three or more copies of each chromosome (triploid or greater organisms), disomy is an aneuploid chromosome condition. In uniparental disomy, both copies of the chromosome come from the same parent (no contribution from the other parent).

The term "trisomy" as used herein refers to the presence of three copies of a particular chromosome rather than two copies. The presence of an extra chromosome 21 found in Down's syndrome in humans is called "trisomy 21". Trisomy 18 and trisomy 13 are two other human autosomal trisomies. Sex chromosomal trisomy may be found in females (eg, triple X syndrome 47, XXX) or males (eg, Klinefelter syndrome 47, XXY; or Jacobs syndrome 47, XYY). In some embodiments, the trisomy is a duplication of most or all autosomes. In certain embodiments, the trisomy is a pan-chromosomal aneuploidy resulting in three instances (eg, three copies) of a particular type of chromosome (eg, two copies of a particular type of chromosome for euploidy). instead of one instance (e.g. pair).

The terms "tetrasomy" and "pentasomy" as used herein refer to the presence of four or five chromosome copies, respectively. Sex chromosome tetrasomies and pentasomies, rarely found on autosomes, have been reported in humans, including XXXX, XXXY, XXYY, XYYY, XXXXXX, XXXXY, XXXYY, XXYYY, and XYYYY.

Chromosomal abnormalities can be caused by various mechanisms. Mechanisms include (i) chromosomal nondisjunction resulting from weakened mitotic checkpoints, (ii) inactive mitotic checkpoints that cause chromosomal nondisjunction on multiple chromosomes, (iii) single mitotic checkpoint Meroteric binding that occurs when the drug substance binds to both mitotic spindle poles, (iv) multipolar spindle formation when more than two spindle poles are formed, (v) single spindle Including, but not limited to, unipolar spindle formation when only poles are formed, and (vi) tetraploid intermediates resulting from the end result of the unipolar spindle machinery.

The terms "partial monosomy" and "partial trisomy" as used herein refer to an imbalance of genetic material caused by the loss or gain of part of a chromosome. Partial monosomy or partial trisomy can result from an unbalanced translocation, in which an individual carries an induced chromosome formed by the break and fusion of two different chromosomes. In this situation, an individual has only three copies of one chromosome part (the two normal copies, plus the segment present on the induced chromosome) and one copy of the other chromosome part contained in the induced chromosome. will have

The term "mosaicism" as used herein refers to chromosomal aneuploidy in some but not all cells of an organism. Certain chromosomal abnormalities can exist as mosaic and non-mosaic chromosomal abnormalities. For example, certain trisomy 21 individuals have mosaic Down syndrome and some have non-mosaic Down syndrome. Different mechanisms may cause mosaicism. For example: (i) the original zygote is thought to have three 21st chromosomes, which usually results in simple trisomy 21, but during the course of cell division, one or more cell lineages develop 21st chromosomes. and (ii) the original zygote appeared to have two chromosomes 21, but during cell division one of the 21st chromosomes was duplicated. Somatic mosaicism may arise through mechanisms different from those commonly associated with genetic syndromes with complete or mosaic chromosomal aneuploidies. Somatic mosaicism has been identified, for example, in certain types of cancer and neurons. In one particular case, trisomy 12 was identified in chronic lymphocytic leukemia (CLL) and trisomy 8 in acute myelogenous leukemia (AML). Also, hereditary syndromes in which individuals are prone to chromosome breaks (chromosomal instability syndromes) are frequently associated with increased risk for various types of cancer, and thus somatic chromosomal aberrations in carcinogenesis. The role of numerology is noted. The methods and protocols described herein can identify the presence or absence of non-mosaic and mosaic chromosomal abnormalities.

Tables 1A and 1B present a non-limiting list of chromosomal conditions, syndromes, and/or abnormalities that may be identified by the methods, machines and/or devices described herein. Table 1B is from the DECIPHER database as of Oct. 6, 2011 (e.g., version 5.1, based on locations mapped to GRCh37; uniform resource locator (URL) dechipher.sanger.ac.uk available).

Grade 1 conditions often have one or more of the following characteristics; pathogenic abnormality; strong agreement among geneticists; high penetrance; but also have some general characteristics; all cases in the literature have clinical manifestations; no cases of healthy individuals with abnormalities; not reported in DVG database or in healthy population Functional data confirming single-gene or multigene quantitative effect; Confirmed or strong candidate gene; Clinical management proposal defined; Cancer risk known and survey proposal; Multiple (OMIM, Gene reviews, Orphanet, Unique, Wikipedia); and/or available for diagnostic use (pregnancy counseling).

Grade 2 conditions often have one or more of the following characteristics; potential pathogenic abnormality; high penetrance; /Low number of reports; all cases reported have clinical manifestation; no functional data or confirmed virulence gene; multiple sources (OMIM, Gene reviews, Orphanet, Unique, Wikipedia) and/or for diagnostic purposes and pregnancy counseling.

Grade 3 conditions often have one or more of the following characteristics; a susceptibility locus; unaffected parents of a healthy individual or proband described; present in a control population; Non-penetrant; mild and nonspecific presentation; less consistent characterization; no functional data or confirmed virulence genes; more limited source of data; deviating from majority or if new clinical findings are present, the possibility of a second diagnosis remains probable; and/or caution when used for diagnostic purposes, and for pregnancy counseling , be careful with advice.

Preeclampsia In some embodiments, the presence or absence of preeclampsia is determined using the methods, machines or devices described herein. Pre-eclampsia is a condition in which hypertension occurs during pregnancy (eg, pregnancy-induced hypertension) and is associated with significant amounts of protein in the urine. In certain embodiments, pre-eclampsia is also associated with increased levels of extracellular nucleic acids and/or altered methylation patterns. For example, a positive correlation was observed between extracellular fetal hypermethylated RASSF1A levels and the severity of pre-eclampsia. In one particular example, there is increased DNA methylation for the H19 gene in preeclamptic placentas compared to normal controls.

Pre-eclampsia is one of the leading causes of maternal and fetal/neonatal mortality and morbidity worldwide. Circulating cell-free nucleic acids in plasma and serum are novel biomarkers with promising clinical applications in different medical fields, including prenatal diagnosis. Changes in cell-free fetal (cff) DNA in maternal plasma, when quantified using real-time quantitative PCR for, for example, the male-specific SRY or DYS14 loci, are indicative for impending pre-eclampsia. reported in different trials. Elevated levels may be seen early in pregnancy in cases of early-onset pre-eclampsia. Elevated levels of preclinical cffDNA may also be due to hypoxia/reoxygenation within the intervillous space causing increased tissue oxidative stress and placental apoptosis and necrosis. In addition to evidence for increased efflux of cffDNA into the maternal circulation, there is also evidence for reduced renal excretion of cffDNA in pre-eclampsia. Since the amount of fetal DNA is currently determined by quantification of Y-chromosome-specific sequences, alternative approaches such as measurement of cell-free total DNA or gender-independent fetal epigenetic markers such as DNA methylation. Use provides an alternative. Cell-free RNA from placenta is another alternative biomarker that can be used to screen and diagnose pre-eclampsia in clinical practice. Fetal RNA is associated with intracellular placental particles that protect it from degradation. Fetal RNA levels can be 10-fold higher in pregnant females with pre-eclampsia compared to controls and are thus an alternative biomarker that can be used to screen and diagnose pre-eclampsia in clinical practice. .

Pathogens In some embodiments, the presence or absence of a disease state is determined by a method, machine or apparatus described herein. A condition can be caused by infection of the host with pathogens including, but not limited to, bacteria, viruses, or fungi. Because pathogens generally have nucleic acids (e.g., genomic DNA, genomic RNA, mRNA) that are distinguishable from host nucleic acids, the methods, machines and devices provided herein determine the presence or absence of pathogens. can be used to Pathogens often have nucleic acids that are unique to the particular pathogen, eg, epigenetic status and/or mutations, duplications, and/or deletions of one or more sequences. Thus, the methods provided herein can be used to identify specific pathogens or pathogen variants (eg, strains).

Cancer In some embodiments, the presence or absence of a cell proliferative disorder (eg, cancer) is determined using the methods, machines or devices described herein. For example, levels of cell-free nucleic acids in serum can be elevated in patients with various types of cancer compared to healthy patients. For example, patients with metastatic disease can have serum DNA levels about twice as high as non-metastatic patients. Patients with metastatic disease may also be identified by cancer-specific markers and/or, for example, certain single nucleotide polymorphisms or short tandem repeats. Non-limiting examples of cancer types that may be positively correlated with increased levels of circulating DNA include breast cancer, colorectal cancer, gastrointestinal cancer, hepatocellular carcinoma, lung cancer, melanoma, non-Hodgkin's lymphoma, leukemia, Multiple myeloma, bladder cancer, hepatoma, cervical cancer, esophageal cancer, pancreatic cancer, and prostate cancer. Various cancers have nucleic acids with characteristics distinguishable from nucleic acids derived from non-cancerous healthy cells, such as epigenetic conditions and/or sequence mutations, duplications, and/or deletions. and sometimes release it into the bloodstream. Such features may be specific to a particular type of cancer, for example. Accordingly, it is further contemplated that the methods provided herein can be used to identify specific types of cancer.

As described in more detail hereinafter, software can be used to perform one or more steps in the processes described herein, including but not limited to; , data processing, generating results, and/or providing one or more proposed recommendations based on the results generated.

Machines, Software and Interfaces Certain processes and methods described herein often cannot be performed without computers, processors, software, modules or other devices. The methods described herein are generally computer-implemented methods in which one or more portions of the method are controlled by one or more processors (e.g., microprocessors), computers, or microprocessors. may be performed by the device. In some embodiments, one or more or all of the processing methods known or described herein (e.g., mapping, data compression, determination of local genomic bias estimates, determination of relationships, comparison, count normalization, read density and/or read density profile generation, PCA, profile adjustment, portion filtering, portion weighting, profile comparison, profile scoring, outcome determination, etc., or combinations thereof ) is performed by a processor, microprocessor, computer in conjunction with memory, and/or by a device controlled by a microprocessor. Embodiments related to methods described herein are generally applicable to the same or related processes performed by instructions in the systems, apparatus, and computer program products described herein. . In some embodiments, the processes and methods described herein (e.g., quantitation, count counting, and/or determining sequence reads, counts, levels, and/or profiles) are automated methods. performed by In some embodiments, one or more of the steps and methods described herein are performed by a processor and/or computer and/or performed in conjunction with memory. In some embodiments, the automated method includes sequence reads, counts, mappings, mapped sequence tags, levels, profiles, normalizations, comparisons, range setting, sorting, adjustment, plotting, results, embedded in machines including software, modules, processors, peripherals, and/or the like that determine conversion and identification. As used herein, software refers to computer readable program instructions that, when executed by a processor, perform computer operations as described herein.

Reads, counts, read densities, and read density profiles of sequences from test subjects (e.g., patients, pregnant females) and/or from reference subjects determine the presence or absence of mutations in genes. can be further analyzed and processed to Sequence reads, counts, levels, and/or profiles are sometimes referred to as "data" or "datasets." In some embodiments, the data or dataset is one or more characteristics or variables (e.g., sequence-based [e.g., GC content, specific nucleotide sequence, etc.], function-specific [e.g., expressed gene , oncogenes, etc.], location-based [genome-specific, chromosome-specific, segment or segment-specific] characteristics or variables, etc. and combinations thereof). In certain embodiments, data or data sets may be organized into matrices having two or more dimensions based on one or more properties or variables. Data organized in a matrix can be organized using any suitable property or variable. Non-limiting examples of data in the matrix include data organized by maternal age, maternal ploidy, and fetal contribution. In certain embodiments, a data set characterized by one or more properties or variables may be processed after counting.

Apparatus (multiple apparatus, also referred to herein as apparatus in plural), software, and interfaces can be used to implement the methods described herein. Using the devices, software and interfaces, the user can use specific information, programs or processes (e.g., mapping sequence reads, processing mapped data, and/or providing results). Options can be entered, requested, queried, or determined and include, for example, performing statistical analysis algorithms, statistical significance algorithms, statistical variance algorithms, comparisons, iteration steps, validation algorithms, and graphical displays. obtain. In some embodiments, datasets can be entered by a user as input information, a user can download one or more datasets via suitable hardware media (e.g., flash drive), and /or a user can transmit a data set from one system to another system for subsequent processing and/or to obtain results (e.g., from a sequencer to a computer system to read sequences). sending sequence read data for mapping; sending mapped sequence data to a computer system for processing and obtaining results and/or reports).

A system typically includes one or more devices. In some embodiments the device is a machine. In some embodiments, the apparatus includes a machine. An apparatus may include one or more of memory, one or more processors, and/or instructions. If the system includes two or more devices, some or all of the devices may be located at the same location, some or all of the devices may be located at different locations, or all devices may be located at one location. and/or all devices may be located in different locations. Where the system includes two or more devices, some or all of the devices may be co-located with the user, some or all of the devices may be co-located with the user, all devices may be and/or all devices may be located in one or more locations different from the user. Devices of the systems described herein can interface with one or more remote computing servers and/or computers (eg, cloud, cloud computing service) in any suitable manner. The term “cloud,” as used herein, refers to a real-time computer capable of performing centralized functions (e.g., the methods described herein) where portions of the function are shared by multiple computers in a network. Refers in part to two or more computers (eg, often multiple computers) connected by a communication network (eg, the Internet). A "cloud" often allows one or more programs (eg, software programs, modules) to run on multiple connected computers at the same time. In some embodiments, the systems and/or devices described herein comprise clouds (eg, cloud servers, cloud computers, cloud computing services). One or more functions of the systems and/or devices described herein may be performed in the cloud. Data and/or information can be transported to and/or from the device and cloud using any suitable method. The term "computer" as used herein refers to an electrically man-made device that includes a microprocessor capable of performing arithmetic and logical operations. A computer sometimes includes instructions, software (eg, modules), memory, a display, one or more peripherals and/or storage media. In some embodiments the machine includes a computer. In some embodiments the machine is a computer. Computers are often interfaced with and/or connected to other computers (eg, the Internet, networks, clouds).

A system sometimes includes a computing device or sequencing device, or a computing device and a sequencing device (ie, a sequencing machine and/or a computer). Sequencing devices are generally configured to receive physical nucleic acids and produce signals corresponding to the nucleotide bases of the nucleic acids. Sequencing devices are often "loaded" with a sample containing nucleic acids, and the nucleic acids of the sample loaded onto the sequencing device are generally subjected to a nucleic acid sequencing process. The term "loading a sequencing device" as used herein refers to contacting a portion of a sequencing device (e.g., flow cell) with a nucleic acid sample, the portion of the sequencing device performing a nucleic acid sequencing process. is configured to receive a sample for performing In some embodiments, the sequencer is loaded with variants of the sample nucleic acid. Variants are sometimes amplified, restricted, or modified by treatment of the sample nucleic acid to render it suitable for sequencing (e.g., by ligation (e.g., by adding adapters to the ends of the sample nucleic acid by ligation)). digestion, etc., or a combination thereof). Sequencing devices are often partially configured to perform appropriate DNA sequencing methods that produce signals (e.g., electronic signals, detector signals, images, etc., or combinations thereof) corresponding to the nucleotide bases of the loaded nucleic acid. is configured to

The one or more signals corresponding to each base of a DNA sequence are often processed and/or converted by appropriate processes to determine base determination (e.g., specific nucleotide bases such as guanine, cytosine, thymine, uracil, adenine, etc.). etc.). A collection of base calls derived from a loaded nucleic acid is often processed and/or aggregated into one or more sequence reads. In embodiments in which multiple sample nucleic acids are sequenced simultaneously (ie, multiplexed), a suitable demultiplexing process can be leveraged to associate specific reads with the sample nucleic acids from which they originated. As described herein, sequence reads can be aligned to a reference genome by appropriate processing, and reads aligned to portions of the reference genome can be counted.

A sequencing device sometimes is associated with and/or includes one or more computing devices in a system. The one or more computing units are sometimes configured to perform one or more of the following processes: generating base calls from sequencer signals, accumulating reads (e.g., generating reads), Demultiplexing reads, aligning reads against a reference genome, counting reads aligned against a portion of the genome in the reference genome, and more. The one or more computing units are sometimes configured to perform one or more of the following additional processes: normalize read counts (e.g., reduce or remove bias), one or Multiple determinations (e.g., fetal fraction, fetal ploidy, fetal sex, fetal chromosome count, outcome, presence or absence of genetic mutations (e.g., fetal chromosomal aneuploidy (e.g., chromosome 13, 18) , and/or determination of the presence or absence of trisomy 21).

In some embodiments, one computing unit is associated with the sequencer, and in certain embodiments, one computing unit performs most or all of the following processing: sequencer signal generating base calls from, accumulating reads, demultiplexing reads, aligning reads against a genomic portion of a reference genome and counting reads aligned against this genomic portion, normalizing read counts, and Generation of one or more outcomes (eg, fetal fraction, presence or absence of mutations in particular genes), and the like. In the latter embodiment, where one computing unit is associated with the sequencing device, the computing unit often has one or more processors (e.g., microprocessors) and one or more processors to perform the processing. It contains a memory with the instructions to be implemented. In some embodiments, one computing unit is a single-core computing device local to the sequencing device (e.g., located at the same location (e.g., same address, same building, same floor, same room, etc.) or It can be a multi-core computing device. In some embodiments, one computing device is incorporated into the sequencing device.

In some embodiments, multiple computing units within the system are associated with the sequencing unit, and a subset of the total processing performed by the system is allocated to, or divided among, specific computing units of the system. may have been. A subset of the total number of processes may be divided among two or more computing units or groups thereof in any suitable combination. In certain embodiments, the generation of base calls from sequencer signals, the accumulation of reads, and the demultiplexing of reads is performed by a first computing unit or groups thereof and Alignment and counting of mapped reads is performed by a second computing unit or groups thereof, and normalization of read counts and providing one or more outcomes is performed by a third computing unit or groups thereof. will be In systems including two or more computing units or groups thereof, each particular computing unit may include memory, one or more processors, or a combination thereof. A multi-computing unit system sometimes includes one or more suitable servers local to the sequencing device, and sometimes one or more suitable servers not local to the sequencing device (e.g., web servers, online servers, application server, remote file server, cloud cloud server (e.g. cloud environment, cloud computing)).

Devices in different system configurations may generate different types of output data. For example, a sequencing device can output base signals, and the base signal output data can be transferred to a computing device that converts the base signal data into base calls. In some embodiments, the base calls are output data from one computing unit and transferred to another computing unit to generate sequence reads. In certain embodiments, the base calls are not output data from a particular instrument, but instead are exploited within the same instrument that received the base signals of the sequencing instrument and generated the sequence reads. In some embodiments, a device receives base signals of a sequencer, generates base calls, sequence reads, demultiplexes sequence reads, and sequences sequence reads against a reference genome. Outputs demultiplexed sequence reads for the sample that can be transferred to another device or group of devices for alignment. In some embodiments, one or a group of devices can output aligned sequence reads mapped to portions of a reference genome (e.g., SAM or BAM files), such The output data is a second computing unit that normalizes sequence reads (e.g., normalizes sequence read counts) and produces an outcome (e.g., fetal fraction and/or presence or absence of fetal trisomy). Or can be transferred to the group. Output data from one device can be transferred to a second device in any suitable manner. For example, output data from one device is sometimes placed on a physical storage device that is transported and connected to a second device to which the output data is transferred. Output data is sometimes stored by one device in a database and a second device accesses the output data from the same database.

The system sometimes includes a bias reducer. A bias reducer sometimes includes one or more computers. In some embodiments, the bias reducer maps sequence reads and/or compresses reads (eg, mapped sequence reads). The bias reducer sometimes compresses array reads into a suitable compressed format (eg, BReads format). In some embodiments, the bias reducer produces lead densities, density profiles, adjusted lead density profiles, and/or outcomes. One or more functions of the bias reducer may be performed by a network and/or cloud (eg, cloud computing network). A bias reducer may comprise a plurality of servers ( For example, it can interface with a cloud server). The bias reducer can transport data and/or information to the cloud, where one or more functions of the bias reducer are performed. Processed data and/or information can be transported from the cloud to the bias reducer.

The system sometimes includes a sequencer and a bias reducer, wherein the sequencer generates sequence reads from the sample nucleic acid, sometimes maps the sequence reads, unmapped sequence reads or mapped sequence reads. is provided and/or transferred to the bias reducer. The sequencer can provide or transfer leads to the bias reducer by any suitable method. The sequencer and bias reducer are sometimes connected together by a suitable hardware interface. In some embodiments, the sequencer and bias reducer are network and/or cloud connected. In some embodiments, the sequencer and bias reducer are connected together by a network and/or cloud. Some or all of the methods and/or functions of the sequencer and/or bias reducer may be performed by the cloud. The sequencing machine can transfer the reads to the bias reduction machine using transitory and/or non-transitory computer readable media. For example, array leads can be transferred by digital or analog signals transferred by wired cables and/or wireless signals. In some embodiments, sequence reads are transferred from the sequencer to the bias reducer using non-transitory computer-readable storage media.

A bias reducer may include one or more modules described herein that may implement some or all of the functions of the bias reducer. In some embodiments, the bias reducer includes a compression module and performs the functions of the compression module. In some embodiments, the bias reducer includes one or more of a bias density module, a relationship module, a bias correction module, and/or a multivariate correction module. A bias corrector can use one or more modules to remove bias (eg, GC bias) from reads and/or provide normalized counts of sample reads. In some embodiments, the bias corrector includes one or more of a distribution module, a filtering module, and/or a profile generation module. A bias corrector can often process sequence reads from a training set or reference and sequence reads from a test sample. In some embodiments, the bias corrector includes one or more of a PCA statistics module and/or a partial weighting module. Bias correctors often utilize mapped reads and multiple modules to provide read densities, density profiles, and/or adjusted read density profiles to scoring modules, end users, computer peripherals (e.g., displays, printer), or to an outcome generator. In some embodiments, the bias reducer provides an outcome. Sometimes bias reducers do not provide outcomes. In some embodiments, bias reducers include outcome generators. Sometimes, the bias reducer transfers normalized reads, read densities, density profiles, and/or adjusted read density profiles to the outcome generator. The bias reducer can transfer data and/or information (eg, read density profiles) to the outcome generator by any suitable method. In some embodiments, the system includes one or more of a sequencer, a bias reducer, and/or an outcome generator. The outcome generator can receive the normalized count of reads, the read density, the density profile, and/or the adjusted read density profile from the bias corrector. Outcome generators often provide a judgment or outcome (eg, determining the presence or absence of a genetic mutation). Outcome generators often provide judgments or outcomes to end users and/or computer peripherals (eg, displays, printers). Outcome generators sometimes implement one or more of a filtering module, a distribution module, a profile generation module, a PCA statistics module, a partial weighting module, a scoring module, and/or one or more other suitable modules. include.

In some embodiments, the user interacts with the device (eg, computing device, sequencing device). In some embodiments, a user can query a system, computer, or module, which in turn can obtain a data set via internet access (e.g., cloud), and a particular In the embodiment of , the programmable processor may be prompted to obtain the appropriate data set based on the given parameters. The programmable processor may also prompt the user to select one or more data set options selected by the processor based on the provided parameters. A programmable processor may prompt a user to select one or more processor-selected data set options based on information found via the Internet, other internal or external information, or the like. The options are one or more data characteristic selections, one or more statistical algorithms, one or more statistical analysis algorithms, one or more statistical significance algorithms, iteration steps, one or more validity A identity verification algorithm and one or more graphical representations of a method, apparatus, or computer program may be selected for selection.

The systems covered by this specification may include common components of computer systems, such as network servers, laptop systems, desktop systems, handheld systems, personal digital assistants, computing kiosks, and the like. A computer system may include one or more input means such as a keyboard, touch screen, mouse, voice recognition means, or other means that allow a user to enter data into the system. The system may include a display screen (e.g., CRT or LCD), speakers, fax machine, printer (e.g., laser, inkjet, impact, black-and-white or color printer), or visual, auditory and/or hard copy of information. It can further include one or more outputs, including, but not limited to, other outputs (eg, outcomes and/or reports) that are useful for providing outputs. In some embodiments, the display module is a suitable visual display for presentation on a suitable display (e.g., monitor, LED, LCD, CRT, etc., or combinations thereof), printer, suitable peripheral or device. Process, transform, and/or transport data and/or information in media. In certain embodiments, the display module provides a visual representation of relationships, profiles, or outcomes. Non-limiting examples of suitable visual media and/or displays include charts, plots, graphs, etc., or combinations thereof. In some embodiments, the display module processes and transforms the data and/or information into a visual representation of the fetal and/or maternal genome or segments thereof (eg, chromosomes or portions thereof). Some embodiments require a display module, or a machine that includes a display module, to provide a suitable visual indication.

In the system, input and output means may be connected to a central processing unit which may include, among other components, a microprocessor for executing program instructions, and memory for storing program codes and data. In some embodiments, the process may be implemented as a single user system located at a single geographic location. In certain embodiments, the process may be implemented as a multi-user system. In a multi-user implementation, multiple central processing units may be connected by a network. A network may be a division within a part of a building, local to span an entire building, span multiple buildings, span an area, span an entire country, or be global. The network may be private, owned and managed by a provider, or implemented as an Internet-based service such that users access web pages to enter and retrieve information. Thus, in certain embodiments, a system includes one or more machines that may be local or remote to a user. More than one machine at one location or multiple locations may be accessed by a user and data may be mapped and/or processed serially and/or in parallel. Accordingly, suitable configurations and control methods are available for mapping and/or processing data using multiple machines, such as in local networks, remote networks, and/or "cloud" computing platforms.

A system, in some embodiments, may include a communication interface. A communications interface allows software and data to be transferred between a computer system and one or more external devices. Non-limiting examples of communication interfaces include modems, network interfaces (such as Ethernet cards), communication ports, PCMCIA slots and cards, and the like. Software and data transferred via communication interfaces generally take the form of signals, which may be electronic, electromagnetic, optical and/or other signals received by the communication interface. Signals are often provided to a communication interface via channels. A channel often carries a signal and may be implemented using wire or cable, fiber optics, telephone wire, cellular telephone link, RF link, and/or other communication channels. Thus, in one example, the communication interface can be used to receive signal information that can be detected by a signal detection module.

Data can be input by any suitable device and/or method including, but not limited to, manual input devices or direct data entry devices (DDE). Non-limiting examples of manual devices include keyboards, concept keyboards, touch-sensitive screens, light pens, mice, trackballs, joysticks, graphics tablets, scanners, digital cameras, video digitizers, and voice recognition devices. Non-limiting examples of DDEs include bar code readers, magnetic strip codes, smart cards, magnetic ink character recognition, optical character recognition, optical mark recognition, and turnaround documents.

In some embodiments, output from a sequencing device can serve as data that can be input via an input device. In certain embodiments, mapped sequence reads can serve as data that can be input via an input device. In certain embodiments, simulation data is generated by an in silico process, and post-simulation data can serve as data that can be input via an input device. The term "in silico" refers to computer-aided research and experimentation. In silico processes include, but are not limited to, mapping sequence reads and processing mapped sequence reads by the processes described herein.

The system can include software useful for performing the processes described herein, and the software can include one or more modules that perform such processes (e.g., sequencing module, bias correction module , display module). The term “software” refers to computer-readable program instructions that, when executed by a computer, perform computer operations. Instructions executable by one or more processors may be provided as executable code that, when executed, can cause one or more processors to perform the methods described herein. The modules described herein may exist as software and instructions (eg, processes, routines, subroutines) embedded in software may be implemented or executed by a processor. For example, a module (eg, software module) may be part of a program that performs a particular process or task. The term "module" refers to a self-contained functional unit that can be used in a larger device or software system. A module may include a set of instructions for implementing the module's functionality by one or more microprocessors. Module instructions can be written in UNIX, Linux, oracle, windows, Ubuntu, ActionScript, C, C++, C#, Haskell, Java, JavaScript (non-limiting examples). (registered trademark), Objective-C, Perl, Python, Ruby, Smalltalk, SQL, Visual Basic, COBOL, Fortran, UML, HTML (e.g. with PHP), PGP, G, R, S, etc., or combinations thereof. implemented within a computing environment by using suitable programming languages, suitable software, and/or code written in suitable languages (e.g., computer programming languages known in the art) and/or operating systems, including can do. In some embodiments, the modules described herein include code (eg, scripts) written in S or R that leverages appropriate packages (eg, S packages, R packages). The Comprehensive ^R Archive Network (CRAN) [online], [2013-04-24 Searched at]), R, R source code, R programs, R packages, and R documentation are available for download. CRAN is a worldwide network of ftp and web servers that store the same up-to-date versions of the code and documentation for R.

A module may transform data and/or information. The data and/or information may be in any suitable form. For example, data and/or information can be digital or analog. In certain embodiments, data and/or information may be packets, bytes, codes, or bits. In some embodiments, data and/or information may be any collected, aggregated, or usable data or information. Non-limiting examples of data and/or information include suitable media, files, pictures, videos, sounds (e.g. frequencies, audible or inaudible), numbers, constants, values, objects, times, documents, functions, instructions , computer code, maps, references, sequences, reads, mapped reads, read densities, read density profiles, ranges, thresholds, displays, presentations, outcomes, transformations, etc., or combinations thereof. A module accepts or receives data and/or information, transforms the data and/or information into a second form, and provides or transfers the second form to a machine, peripheral, component, or another module. can be done. A module can perform one or more of the following non-limiting functions: e.g. perform PCA; provide principal components; adjust read densities and/or read density profiles; weight fractions; score; providing estimates of local genomic bias, providing estimates of local genomic bias, providing bias frequencies, providing read densities, providing read density profiles, decision ranges and/or decisions Provide no bounds, provide a measure of uncertainty, provide or determine expected ranges (eg, threshold ranges and threshold levels), and/or determine outcomes. A processor, in certain embodiments, may implement instructions within modules. In some embodiments, one or more processors are required to implement instructions within a module or groups of modules. A module may provide data and/or information to another module, device, or source, and may receive data and/or information from another module, device, or source.

A non-transitory computer-readable storage medium sometimes includes an executable program stored therein, and sometimes the program directs a microprocessor to perform a function (e.g., a method described herein). A computer program product may be tangibly embodied in tangible computer-readable media or tangibly embodied in non-transitory computer-readable media. A module may be stored on a computer readable medium (eg, disk, drive) or in memory (eg, random access memory). A module and a processor capable of implementing instructions from the module may reside within a machine or different apparatus. A module and/or processor capable of implementing the instructions for the module may be located at the same location as the user (eg, local network) or at a location different from the user (eg, remote network, cloud system). In embodiments in which the method is practiced in conjunction with two or more modules, the modules may reside within the same device and one or more modules may be in the same physical location. It may reside in one different device, and one or more modules may reside in different devices in different physical locations.

The machine, in some embodiments, includes at least one processor that implements the instructions in the modules. A sequence read count mapped to a portion of a reference genome may be accessed from a processor executing instructions configured to perform the methods described herein. The count number accessed by the processor may be in the system's memory, and the count number can be accessed and placed in the system's memory after it is obtained. In some embodiments, the machine includes a processor (e.g., one or more processors) that performs one or more instructions (e.g., processes, routines, and/or subroutines) from the modules and/or can be implemented. In some embodiments, the machine includes multiple processors, such as processors with parallel synchronized operation. In some embodiments, the machine operates with one or more external processors (eg, internal or external networks, servers, storage devices, and/or storage networks (eg, cloud)). In some embodiments, the machine includes modules. In certain embodiments, a machine includes one or more modules. Machines that include modules are often capable of receiving and transferring one or more data and/or information to other modules. In certain embodiments, a machine includes peripherals and/or components. In certain embodiments, the machine includes one or more peripherals or components that can transfer data and/or information to and from other modules, peripherals, and/or components. can contain. In certain embodiments, the machine interacts with peripherals and/or components that provide data and/or information. In certain embodiments, peripherals and components assist the machine in performing certain functions or interact directly with the modules. Non-limiting examples of peripherals and/or components include suitable computer peripherals, I/O or storage methods or devices, including scanners, printers, displays (e.g. monitors, LEDs, LCTs, or CRT), camera, microphone, pad (e.g. ipad, tablet), touch screen, smart phone, mobile phone, USB I/O device, USB mass storage device, keyboard, computer mouse, digital pen, modem, hard drive, jump Drives, flash drives, processors, servers, CDs, DVDs, graphics cards, specialized I/O devices (e.g. sequencers, photocells, photomultipliers, optical readouts, sensors, etc.), one or more flow cells, fluidics including, but not limited to, handling components, network interface controllers, ROM, RAM, wireless transfer methods and devices (Bluetooth®, WiFi, etc.), World Wide Web (www), Internet, computer and/or other modules. Not limited.

Software is often provided on a program product containing program instructions recorded on computer-readable media (e.g., non-transitory computer-readable media); such media include floppy disks, , hard disks, and magnetic tapes; and optical media, including CD-ROM discs, DVD discs, magneto-optical discs, solid state drives, flash drives, RAM, ROM, BUS, floppy disks, etc. and other such media on which program instructions are recordable. When conducted online, servers and websites maintained by the organization may be configured to provide software downloads to remote users, or remote users may access remote systems maintained by the organization to Software can be accessed remotely. The software can obtain or receive input information. The software can include modules that specifically acquire or receive data (e.g., data receiving modules that receive sequence read data and/or mapped read data) and modules that specifically process data (e.g., A processing module that processes the received data (eg, filters, normalizes, provides results and/or reports) may be included. The terms "obtain" and "receiving" input information refer to data (e.g., sequence , mapped reads). Input information may be generated at the same location where it is received, or it may be generated at a different location and transferred to the receiving location. In some embodiments, the input information is modified (eg, placed in a manageable format (eg, tabular)) before being processed.

In certain embodiments, software may include one or more algorithms. Algorithms can be used to process data and/or obtain results or reports with a finite sequence of instructions. An algorithm is often a list of prescribed instructions for completing a task. Starting from an initial state, an instruction may describe an operation that progresses through a defined series of successive states and finally ends with a final ending state. Transitions from one state to the next are not necessarily deterministic (eg, some algorithms incorporate randomness). Examples include, but are not limited to, search algorithms, sorting algorithms, integration algorithms, numerical algorithms, graph algorithms, string algorithms, modeling algorithms, computational geometry algorithms, combinatorial algorithms, machine learning algorithms, cryptography algorithms, data compression algorithms. , parsing algorithms, and the like. The algorithms may include one algorithm or two or more algorithms working in combination. The algorithms may be of any suitable complexity class and/or parameterized complexity. Algorithms can be used to compute and/or process data, and in some embodiments can be used in deterministic or probabilistic/predictive approaches. The algorithms can be implemented within a computing environment using any suitable programming language, non-limiting examples of such languages include C, C++, Java, Perl, R, S, Python , Fortran, etc. In some embodiments, the algorithm uses tolerances, statistical analysis, statistical significance, measures of uncertainty, and/or comparison with other information or data sets (e.g., using neural nets or clustering algorithms). where applicable).

In certain embodiments, several algorithms may be implemented for use within software. These algorithms can be trained using raw data in some embodiments. For each new raw data sample, the trained algorithm can generate a representative processed data set or results. The processed dataset may also have reduced complexity compared to the parent processed dataset. Based on the processed set, in some embodiments, the performance of trained algorithms based on sensitivity and specificity can be evaluated. Algorithms with the highest sensitivity and/or specificity can be identified and utilized in certain embodiments.

In certain embodiments, simulated (or simulated) data can assist data processing, for example, by training an algorithm or testing an algorithm. In some embodiments, the simulated data includes hypothetical different samplings of different groupings of sequence reads. With simulated data, what can be expected from the true population or what can be skewed in testing algorithms and/or assigning correct classifications can be the basis. Simulated data is also referred to herein as "virtual" data. A simulation may be performed by a computer program in certain embodiments. One possible step in using simulated data sets is to assess the confidence of the identified results, e.g. how well the random sampling matches the original data, or how well the original data to express or evaluate. One approach is to calculate a probability value (p-value), which estimates the probability that a random sample will score better than a selected sample. In some embodiments, an empirical model may be evaluated, which assumes that at least one sample matches the reference sample (with or without degradation mutations). In some embodiments, another distribution, such as the Poisson distribution, can be used to define the probability distribution.

A system may include one or more processors in certain embodiments. A processor may be connected with a communication bus. A computer system may include main memory, often random access memory (RAM), and may also include secondary memory. In some embodiments, memory includes non-transitory computer-readable storage media. Secondary memory includes, for example, hard disk drives and/or removable storage drives, which may include floppy disk drives, magnetic tape drives, optical disk drives, memory cards, and the like. Removable storage drives often read from and/or write to removable storage units. Non-limiting examples of removable storage units include floppy disks, magnetic tapes, optical disks, etc., which are readable and writable by eg removable storage drives. Removable storage units may include computer usable storage media that contain computer software and/or data.

The processor can implement software within the system. In some embodiments, the processor can be programmed to automatically perform tasks described herein that can be performed by a user. As such, a processor or an algorithm implemented by such a processor may require little or no oversight or input by a user (eg, software may be programmed to automatically perform functions). In some embodiments, the process is too complex to determine the presence or absence of genetic mutations in a single individual or group of individuals within a short enough time frame. It is impossible to do the process in

In some embodiments, the secondary memory may include other similar means for allowing computer programs or other instructions to be loaded into the computer system. For example, a system may include removable storage units and interface devices. Non-limiting examples of such systems include program cartridges and cartridge interfaces (such as those found in video game devices), removable memory chips (such as EPROM or PROM), and associated sockets, and software and data transfer from removable storage units. Other removable storage units and interfaces are included that allow it to be moved to a computer system.

In some embodiments, an entity is generating a sequence read count, mapping the sequence reads to the portion, counting the mapped reads, and counting the read counts. Subsequent mapped reads can be utilized in the methods, systems, machines, or computer program products described herein. In certain embodiments, the sequence read count mapped to the portion is used by a second entity in a method, system, machine, or computer program product described herein to: It may also be moved by one entity to a second entity.

In some embodiments, one entity generates the sequence reads, and in some embodiments, a second entity maps the sequence reads to portions within the reference genome. A second entity may count the mapped reads and utilize the counted mapped reads in a method, system, machine, or computer program product described herein. In certain embodiments, the second entity transfers the mapped reads to a third entity, the third entity counts the mapped reads, and the mapped reads are referred to herein as in any method, system, machine, or computer program product described in the In one particular embodiment, the second entity counts the mapped reads and transfers the counted mapped reads to a third entity, the third entity counting the Subsequent mapped reads are utilized in the methods, systems, machines, or computer program products described herein. In embodiments involving a third entity, the third entity may be the same as the first entity. That is, a first entity may transfer sequence reads to a second entity that maps and/or maps the sequence reads to portions within the reference genome. The read can be counted, and the second entity can transfer the mapped and/or counted read to the third entity. A third entity may be utilized in a method, system, machine, or computer program product described herein for mapped and/or counted leads, in which case the third The entity may be the same as the first entity, or the third entity may be different from the first or second entity.

In some embodiments, one entity obtains blood from a pregnant female, optionally isolates nucleic acid from the blood (e.g., from plasma or serum), and a second entity generates sequence reads from the nucleic acid. transfer blood or nucleic acids to another entity;

FIG. 11 illustrates a non-limiting example computing environment 510 in which the various systems, methods, algorithms, and data structures described herein can be implemented. Computing environment 510 is only one example of a suitable computing environment and is not intended to suggest any limitation as to the scope of use or functionality of the systems, methods and data structures described herein. Moreover, computing environment 510 should not be interpreted as having any dependency or requirement relating to any one or combination of components illustrated in computing environment 510 . A subset of the systems, methods and data structures shown in FIG. 11 may be used in certain embodiments. The systems, methods and data structures described herein are operational with numerous other general purpose or special purpose computing system environments or configurations. Examples of known computing systems, environments, and/or configurations that may be suitable include personal computers, server computers, thin clients, thick clients, portable or laptop devices, multiprocessor systems, microprocessor-based systems, and sets. Top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, and the like.

Operating environment 510 of FIG. 11 includes a general-purpose computing device in the form of computer 520 having processing unit 521 , system memory 522 , and various system components including system memory 522 operatively coupled to processing unit 521 . A system bus 523 is included. There may be only one processing unit 521, or there may be more than one, such that the processor of computer 520 includes a single central processing unit (CPU) or multiple processing units commonly referred to as a parallel processing environment. obtain. Computer 520 may be a conventional computer, a distributed computer, or any other type of computer.

System bus 523 can be any of several types of bus structures, including memory buses or memory controllers, peripheral buses, and local buses using any of a variety of bus architectures. The system memory, also referred to simply as memory, includes read-only memory (ROM) 524 and random access memory (RAM). A basic input/output system (BIOS) 526 , containing the basic routines that help to transfer information between elements within computer 520 , such as during start-up, is stored in ROM 524 . Computer 520 has a hard disk drive interface 527, not shown, that reads from and writes to a hard disk, a magnetic disk drive 528 that reads from and writes to a removable magnetic disk 529, and a removable optical disk 531, such as a CD ROM or other optical media. It may further include an optical disc drive 530 that reads from and writes to.

Hard disk drive 527, magnetic disk drive 528, and optical disk drive 530 are connected to system bus 523 by hard disk drive interface 532, magnetic disk drive interface 533, and optical disk drive interface 534, respectively. The drives and their associated computer-readable media provide nonvolatile storage of computer readable instructions, data structures, program modules and other data for computer 520 . Any type of computer-readable media capable of storing computer-accessible data, such as magnetic cassettes, flash memory cards, digital video discs, Bernoulli cartridges, random access memory (RAM), read-only memory (ROM), etc. , can be used within the operating environment.

A number of program modules reside on the hard disk, magnetic disk 529, optical disk 531, ROM 524, or RAM, including an operating system 535, one or more application programs 536, other program modules 537, and program data 538. can be stored. A user may enter commands and information into personal computer 520 through input devices, such as keyboard 540 and pointing device 542 . Other input devices (not shown) can include microphones, joysticks, gamepads, satellite dishes, scanners, and the like. These and other input devices are often connected to the processing unit 521 via a serial port interface 546 coupled to the system bus, although other interfaces such as a parallel port, game port, or universal serial bus ( USB) may be connected. A monitor 547 or other type of display device is also connected to system bus 523 via an interface, such as video adapter 548 . In addition to monitors, computers typically include other peripheral output devices (not shown), such as speakers and printers.

Computer 520 can operate in a networked environment using logical connections to one or more remote computers, such as remote computer 549 . These logical connections may be accomplished through a communications device in communication with computer 520, or portions thereof, or otherwise. Although only memory storage device 550 is shown in FIG. 11, remote computer 549 can be another computer, server, router, network PC, client, peer device, or other general network node and associated with computer 520. and generally include many or all of the above elements. 11. The logical connections depicted in FIG. 11 include a local area network (LAN) 551 and a wide area network (WAN) 552 . Such networking environments are commonplace in office networks, enterprise-wide computer networks, intranets and the Internet, all of which are exemplary networks.

When used in a LAN-network environment, the computer 520 is connected to a local network 551, which is a type of communication device, via a network interface or adapter 553. When used in a WAN-network environment, computer 520 often includes a modem 554 , one type of communication device, or any other type of communication device for establishing communications across wide area network 552 . Modem 554 , which may be internal or external, is connected to system bus 523 via serial port interface 546 . In a networked environment, program modules depicted relative to personal computer 520, or portions thereof, may be stored in the remote memory storage device. It will be appreciated that the network connections shown are non-limiting examples and other communication devices for establishing a communications link between computers can be used.

In some embodiments, the system includes one or more microprocessors and memory, wherein the memory contains instructions executable by one or more microprocessors. generates a relationship between (a) (i) local genomic bias estimates and (ii) bias frequencies for sequence reads of a test sample, thereby generating a sample bias relation ( The sequence reads are of circulating cell-free nucleic acids derived from the test sample and mapped against the reference genome); (the reference bias relation is between (i) the local genomic bias estimate and (ii) the bias frequency for the reference), (c) according to the comparison determined in (b) It is configured to normalize the sequence read counts for the sample, thereby reducing sequence read bias for the sample.

In some embodiments, the system includes one or more microprocessors and memory, the memory containing instructions executable by one or more microprocessors. (a) generates a relationship between (i) guanine and cytosine (GC) densities and (ii) GC density frequencies for sequence reads of a test sample, thereby generating a sample GC density relationship (sequence reads are of circulating cell-free nucleic acid derived from the test sample and mapped against the reference genome), (b) comparing the sample GC density relationship with the reference GC density relationship, thereby comparing (the reference GC density relationship is between (i) the GC density and (ii) the GC density frequency for the reference), and (c) for the sequence for the sample according to the comparison determined in (b). It is configured to normalize the read counts, thereby reducing sequence read bias with respect to the sample.

In some embodiments, the system includes one or more microprocessors and memory, the memory containing instructions executable by one or more microprocessors. (a) Filter a portion of the reference genome according to the read density distribution, thereby providing a read density profile for the test sample comprising the read density of the filtered portion (the read density relative to the reference genome A read density distribution is determined for partial read densities for multiple samples. ), (b) using a microprocessor to adjust the read density profile for the test sample according to one or more principal components obtained from a set of known euploid samples by principal component analysis, thereby adjusting (c) comparing the test sample profile to a reference profile thereby providing a comparison; (d) the presence or absence of a chromosomal aneuploidy for the test sample according to the comparison is configured to determine

In some embodiments, a non-transitory computer-readable storage medium containing self-stored executable programs is presented herein. In some embodiments, a non-transitory computer-readable storage medium containing a self-stored executable program comprises a computer program product. In some embodiments, a non-transitory computer-readable storage medium containing self-stored executable programs refers to software. Computer program products are often software. In some embodiments, a non-transitory computer readable storage medium containing a self-stored executable program, the program instructing a microprocessor to: (a) read an array of test samples; Generate a relationship between guanine and cytosine (GC) density and (ii) GC density frequency, thereby generating a sample GC density relationship (sequence reads are of circulating cell-free nucleic acid derived from test sample , mapped against the reference genome); and (ii) the GC density frequency), (c) normalizing the sequence read counts for the sample according to the comparison determined in (b), thereby reducing the sequence read bias for the sample Presented herein is a non-transitory computer-readable storage medium that directs you to do things that are reduced.

In some embodiments, a non-transitory computer-readable storage medium comprising a self-stored executable program, the program instructing a microprocessor to: (a) filter portions of a reference genome according to a read density distribution; , thereby providing a read density profile for the test sample comprising the read densities of the filtered portion (read densities are read densities of sequences of circulating cell-free nucleic acids derived from test samples derived from pregnant females). wherein the read density distribution is determined for partial read densities for a plurality of samples), (b) testing according to one or more principal components obtained from a series of known euploid samples by principal component analysis adjusting a read density profile for the sample, thereby providing a test sample profile comprising the adjusted read density; (c) comparing the test sample profile to the reference profile, thereby providing a comparison; (d) comparing Also provided herein is a non-transitory computer-readable storage medium for directing determining the presence or absence of a chromosomal aneuploidy for a test sample according to the method.

Modules One or more modules can be utilized in the methods described herein, non-limiting examples of modules include compression module, sequencing module, mapping module, filtering module, biased density module, A relationship module, a bias correction module, a multivariate correction module, a distribution module, a profile generation module, a PCA statistics module, a partial weighting module, a scoring module, an outcome module, a display module, etc., or combinations thereof. In some embodiments, a module is a non-transitory computer-readable medium (e.g., computer program product, e.g., software, program) that includes a set of instructions, the set of instructions being executed by one or more microprocessors. to perform a function. In some embodiments, modules include instructions in the form of suitable computer code (eg, source code). Source code sometimes includes programs. Computer code sometimes includes one or more files (eg, text files). The computer code may be stored on any suitable non-transitory storage medium (eg, memory on a computer's hard disk). Computer code files are often arranged into directory trees (eg, source trees). The computer code for the modules may be written in any suitable programming language, non-limiting examples of which include the C programming language, Basic, R, R++, S, java, HTML, etc., or combinations thereof. . In some embodiments, a suitable main program acts as an interpreter of computer code. In some embodiments, a module includes and/or has access to memory. A module may be managed by a microprocessor. In certain embodiments, a module, or a machine containing one or more modules, collects, aggregates data and/or information to or from another module, machine, component, peripheral, or operator of the machine. , receive, obtain, access, retrieve, provide and/or transfer. In some embodiments, data and/or information (e.g., sequence reads, counts, etc.) is provided to the module by a machine that includes one or more of the following: one or more flow cells, cameras, detectors (e.g. photodetectors, photocells, electrical detectors (e.g. amplitude modulated detectors, frequency and phase modulated detectors, phase locked loop detectors), counters, sensors (e.g. pressure, temperature, volume, flow, weight sensors), fluid handling devices, printers, displays (e.g., LEDs, LCTs, or CRTs), etc., or a combination thereof, etc. The machine operator inputs constants, thresholds, formulas, or pre-determined values to the module. A module is often configured to transfer data and/or information to or from another module or machine A module receives data and/or information from another module Non-limiting examples of other modules that can be received include compression module, sequencing module, mapping module, filtering module, bias density module, relation module, bias correction module, multivariate correction module, distribution module, Including a profile generation module, a PCA statistics module, a partial weighting module, a scoring module, an outcome module, a display module, etc., or combinations thereof The modules are capable of manipulating and/or transforming data and/or information. The data and/or information derived or transformed by the module can be transferred to another suitable machine and/or module, non-limiting examples of which include a compression module , Sequencing Module, Mapping Module, Filtering Module, Bias Density Module, Relationship Module, Bias Correction Module, Multivariate Correction Module, Distribution Module, Profile Generation Module, PCA Statistics Module, Partial Weighting Module, Scoring Module, Outcomes Module, Display modules, etc., or combinations thereof Machines containing modules may contain at least one processor In some embodiments, data and/or information is received and/or provided by a machine containing modules Machines, including modules, processors (eg, one or more processors) that perform and/or implement one or more instructions (eg, processes, routines, and/or subroutines) of the modules be able to. In some embodiments, modules operate with one or more external processors (eg, internal or external networks, servers, storage devices, and/or storage networks (eg, cloud)). In some embodiments, the system (eg, the system embodiment shown in FIG. 10) includes a compression module, a sequencing module, a mapping module, a filtering module, a biased density module, a relational module, a bias correction module, a multivariate correction distribution module, profile generation module, PCA statistics module, partial weighting module, scoring module, outcome module, display module, etc., or combinations thereof.

Transformations As noted above, data may also be transformed from one form to another. The terms "converted", "conversion", and grammatical derivatives or equivalents thereof, as used herein, refer to the physical starting material (e.g., nucleic acid of the test and/or reference sample) to a digital representation of the physical starting material (e.g., sequence read data), and in some embodiments, to one or more numerical values that can be used to provide results; or Including further conversion of digital representations to graphical representations. In certain embodiments, one or more numerical and/or graphical representations of digitally represented data can be used to represent the physical genomic context of a test subject (e.g., genomic insertion, virtually and/or visually representing the presence or absence of duplications or deletions; representing the presence or absence of variations in sequence quantities associated with a medical condition). The virtual representation may be further transformed into one or more numerical values, or a graphical representation of the digital representation of the starting material. These methods can convert a physical starting material into a numerical or graphical representation, or into a physical contextual representation of the tested genome.

In some embodiments, the methods and systems herein combine a mixture of multiple polynucleotide fragments found in the blood of pregnant females with specific microbes present within fetal, maternal, or placental cells. Transform into one or more representations of visual structures and/or specific microscopic structures (eg, chromosomes or segments thereof). These polynucleotide fragments are generally different cells and tissues (e.g., maternal, placental, fetal, e.g., muscle, heart, liver, lymphocytes, tumors), different chromosomes, and different genetic elements and/or locations ( For example, centromeric regions, repetitive elements, GC-rich regions, hypervariable regions, different genes, different regulatory elements, introns, exons, etc.). In some embodiments, the systems described herein convert polynucleotide fragments into sequence reads by using a sequencer. In some embodiments, the systems described herein convert biased sequence reads into normalized sequence counts, read densities, and/or profiles. Sequence reads are often converted to normalized sequence counts, read densities, and/or profiles, where the bias is often a bias reducer, and/or one or more appropriate treatments and/or It is significantly reduced by using modules (eg, Mapping Module, Bias Density Module, Relationship Module, Bias Correction Module, and/or Multivariate Correction Module). Read densities and/or read density profiles generated from normalized sequence reads and reduced bias normalized sequence reads are useful for generating more confident outcomes. Sequence reads are often modified by transformations that vary specific sequence read parameters to reduce bias, thereby sometimes providing normalized sequence reads that are transformed into profiles and outcomes.

In some embodiments, transforming the dataset reduces the complexity of the data and/or the dimensionality of the data, thereby facilitating the presentation of results. Dataset complexity may also be reduced during the process of converting physical starting materials into virtual representations of the starting materials (eg, reading sequences representing physical starting materials). Suitable properties or variables can be used to reduce the complexity and/or dimensionality of the dataset. Non-limiting examples of properties that can be selected for use as target properties for data processing include GC content, fetal sex prediction, chromosomal aneuploidy identification, specific gene or protein identification, cancer, disease, inherited gene/trait, identification of chromosomal abnormality, biological category, chemical category, biochemical category, gene or protein category, gene ontology, protein ontology, co-regulated gene, cell signaling gene , cell cycle genes, proteins related to the above genes, gene variants, protein variants, co-regulated genes, co-regulated proteins, amino acid sequences, nucleotide sequences, protein structural data, etc., and combinations of the above. Non-limiting examples of dataset complexity and/or dimensionality reduction; reducing multiple sequence reads to profile plots, reducing multiple sequence reads to numerical values (e.g., values, Z normalization of scores, p-values); reduction of multiple analysis methods to probability plots or single points; principal component analysis of derived quantities, etc., or a combination thereof.

The following examples are presented by way of illustration only and are non-limiting. Accordingly, the following examples describe certain embodiments and are not intended to be limiting of the present technology. Those of skill in the art will readily recognize a variety of non-critical parameters that could be changed or modified to yield essentially the same or similar results.

(Example 1)
Ch AI
ChAI is an exemplary system for determining the presence or absence of a chromosomal aneuploidy in a fetus from sequence reads obtained from a test subject (eg, a pregnant female). An example of a system flow diagram for ChAI is shown in FIGS. 10A and 10B. Sequence reads were obtained from a pregnant female test subject and one or more reference subjects, sometimes referred to herein as a training set. Pregnant female subjects in the training set had fetuses confirmed to be euploid by other testing methods.

Sequence reads were first compressed from SAM or BAM format to binary read format (BRead format) to allow ChAI to run more quickly. The BRead format preserves the genomic location and discards other information for each read, including the base pair location determined by the chromosome and reference genome. A BRead file starts with a count of the reads it contains. This eliminates the need for memory relocation and improves loading time. Values were stored on disk as a 4-byte array. The reads are then stored using a 5-byte format, one for the chromosomal ordinal number (1-22, zero-index for X, Y, M) and four for the chromosomal location. The BRead file was loaded by first reading the read count number of the sequence from the first 4 bytes. Each sequence read is then loaded 5 bytes at a time, the first byte indicating the chromosomal ordinal number and the next 4 bytes converting to an integer position. Random sampling of reads can be done quickly by using a disk-skip command for a particular read index.

As an example, for mapped read, 17,673,732, the disk usage of different formats is compared in Table I with the disk usage of BRead format.

The BRead format is about 1/50 smaller than the original SAM file and its space usage is about 12% less than the GZip format. BRead also has the advantage of storing the number of reads at the beginning of a one-time memory allocation, and since reads do not have to be read in order, they can be sampled quickly. These properties were not possible in other formats.

GC Bias Modeling A GC bias model was then acquired for each sample. Samples designated for training were partially used to create partial filters and to learn other genomic biases that are not fully explained by GC bias alone. Finally, training statistics were used to filter and score test samples.

ChAI modeled GC bias using density estimates of local GC content. GC densities were estimated from the reference genome using kernel functions such as the Epanechnikov kernel (Fig. 1). Other kernels are also suitable, including Gaussian or triweight kernels. The bandwidth was chosen as 200 bp, but the bandwidth parameter is flexible.

Kernels were used to estimate GC density at base pair resolution for the reference genome (eg, as shown in Figure 2). A reference GC density estimate was used to determine the local GC content of each read from the sample. The distribution of GC density estimates for the samples was then compared to the distribution across the reference genome to determine GC bias (Fig. 3). Reads and references mapping to AT-rich regions (GC density=0) were discarded.

The difference between the GC density distribution of the sample and that of the reference was modeled using a polynomial fit to the log ratio of the reference distribution density divided by the sample distribution density (Fig. 4). The model was fitted in a weighted fashion, where each weight was taken from the distribution density value of the sample for a given GC density value. This ensured that the tails of the distribution did not overfit. Other fitting models, such as quantile regression analysis models or parameterized probability distributions, can be used as well as suitable distributions of bias.

To adjust for cases where a sample was over- or under-represented compared to the reference, a GC-fit model was used to weight each count of sequence reads for the sample. By incorporating these weights into the read density estimates, the ChAI algorithm was able to correct for GC bias.

Multidimensional Bias Correction GC bias was just one of several biases affecting read patterns within the genome. Additional biases were sometimes modeled and corrected for lead weight estimation using generalized multivariate models. This correction was done as follows:

1. N-bias values were estimated for the test samples and the reference genome for each of the subsets of genomic locations.

2. An N-dimensional smoothing kernel or a suitable parametric function was used to model the density of bias values.

3. A log ratio was calculated for a series of density values obtained from the reference and test densities.

4. Using the selected points, a multivariate model was used to model the log ratio of the densities (eg, a weighted cubic polynomial for each dimension).

5. The model was used to estimate the ratio of frequencies for a given lead compared to the reference and assigned an appropriate weight.

Partial Filtering Samples were scored for chromosomal aberrations based on representation of sequence reads (eg counts) on the genome. This representation was determined using a density function similar to that used for local GC estimation. Read density kernels generally have much larger bandwidths, defaulting to 50,000 bp. Each count of a read contributes to a density equal to its weight derived from the GC's bias model. Read densities can be evaluated at any or all base pairs, but for computational performance reasons only certain locations were used. This position was called a "portion". The part can be located wherever it is most important for estimating lead density. When classifying chromosomal aneuploidies, the parts are initially (eg, before filtering) evenly spaced across the genome. Each portion consisted of a 50,000 bp window that overlapped the next adjacent portion by 25,000 bp before filtering.

Some parts contained poorly mapped genomic regions, causing extreme perturbations in read density from sample to sample. ChAI identified and removed this part through a filtering process using the training set. Portions that showed large deviations in median values (eg, FIG. 5A) and/or MAD values (eg, FIG. 5B) were removed from the study. This deviation threshold was set as any value lying outside the quartiles of the training population and more than four times the interquartile range (Fig. 5). This threshold can be fine-tuned to maximize test performance for a particular set of ChAI parameters.

Training and Scoring Only the reads mapping to the filtered portion were used to calculate the genomic read density profile for each sample. The samples that were part of the training set were then used to estimate the training statistic, which was used to score the test set. This statistic consisted of partial medians, principal components, and null distributions for scoring test statistics. Partial medians and principal components were used to model the genome-wide read bias that can exist from any number of biological and technical artifacts (FIGS. 6A-C). To minimize the impact of extreme partial values on other samples, each value that spanned other parts in the sample and outside 4×IQR was truncated to 4×IQR.

Test samples were corrected for hidden bias by subtracting the initially trained median value from the test portion values. We also removed the sample-valued components that correlated with the top-trained principal component. This was done by modeling the partial values using multivariate linear regression based on principal component terms (Fig. 7). The model-predicted values were subtracted from the sample values, leaving only the unbiased residuals. The number of principal components used is optional and defaults to eight.

After correction, samples were scored using Fisher's exact test. This test compared the number of segments with values greater or less than the trained median within the chromosomal region of interest. This count was evaluated against the rest of the genome. Scoring statistics were set as -log ₁₀ (p-value). Other scoring statistics such as the Wilcoxon signed-rank test or the F-test can also be used at this step.

The test statistic increased in both training and test samples due to residual correlations between parts. This increase was estimated from bootstrapping the training set (Fig. 8).

Scores for test samples were corrected using this null distribution as the experimental background. Scores that were much larger than those in the experimental distribution were corrected using Pareto extrapolation for the tail of the null distribution.

Gender Determination Gender was determined from the principal component profiles of the samples. In the training data set, the second principal component (eg PC2) was highly correlated with gender. Using the regression coefficient of this component as the test statistic, it was a very accurate gender test (Figs. 9A-9B).

Removal of Partial Dependence To improve the predictive power of this approach, additional steps were performed during ChAI actuation. This involved reducing the amount of correlated structure within the part-sample matrix, which better supported the test hypothesis of variable independence, and suppressed the significant score frequency within the null permutation. The approach involved replacing the parts with orthogonal unique parts containing nearly all the same information but no correlation structure.
The first step was to master the transformation matrix Meig for a set of training parts M:
1. SVD decomposition: M=U*D*VT
2. Choose the number of independent unique parts N: (eg, such that the integrated fraction of the N diagonal elements of D is greater than 95%)
3. Compute generalized inverse: Meig=pinv(U[...,1:N]*D[1:N,1:N])

For any subset of the submatrix M, left multiplication by its corresponding Meig resulted in a reduced-dimensional, uncorrelated representation for that subset. Thus, Meig was derived based on the training dataset and applied to the test samples without further modification.

Meig was also used in transforming test variables. Test variables were represented as vectors consisting of all zeros, with zeros placed at expected deviations (eg Chr21 portion). This vector was transformed using Meig with left multiplication so that the transformed partial data were properly matched.

This approach can construct at most as many independent eigenparts as there are samples in the training set. For example, with a training set of 50,000 parts and 1,000 samples, the transformed data contains at most 1,000 parts. This can lead to overcompensation and greatly reduce the number of parts. This approach can be made more relaxed by computing separate Meig transforms and applying them separately for smaller subsets of the partial data. This was particularly useful for removing local correlation structures from neighboring parts.

Other approaches can also be used to reduce the correlation structure of the part. For example, many clustering methods can be used to group parts and replace them with a smaller set of aggregated parts (eg, based on group means or centroids).

(Example 2)
Distribution/Profile Generation Module A script was written in java format to generate a read density profile from sequence read data (eg BRead). The code below is designed to collect read data for each sequence read and also update the density profile with a suitable read density window (e.g., individual read densities for the fraction), fractional median or midpoint Reads were weighted by their distance from and based on sample GC bias correction (see Example 4). The script below can determine or utilize the weighted counts and/or normalized counts generated from the relevance module or the bias correction module (Example 4). In some embodiments, the distribution module may include some or all of the java script shown below, or variations thereof. In some embodiments, the profile generation module may include some or all of the following java scripts, and variations thereof:

(Example 3)
Filtering Module A script was written in R format to filter parts of the read density profile. This code examines the read density profile across samples and identifies fractions that are retained and/or discarded (eg, removed from analysis) based on interquartile range. In some embodiments, the filtering module includes some or all of the following R scripts, and variations thereof:

(Example 4)
Bias Density Module, Association Module, Bias Correction Module and Plotting Module A script was written in R format to generate bias densities, generate and compare associations, and correct for bias within sequence reads. The code is configured for each sample and reference to analyze one or more samples based on local genomic bias estimates (e.g., GC density) and bias models (e.g., association and/or or relevance comparison). The script below directs one or more processors to: (i) guanine and cytosine (GC) densities and (ii) GC densities for reading sequences of test samples, to: (b) comparing the GC density relationship of the sample with the GC density relationship of the reference, thereby generating a comparison; but in this case the GC density relevance of the reference is the relationship between (i) the GC density and (ii) the GC density frequency with respect to the reference, including the appropriate modification of the script, (c)(b ) normalizes the sequence read counts for the sample, but in this case the sequence read bias for the sample is reduced. In some embodiments, the Biased Density Module, Relevance Module, Bias Correction Module, and/or Plotting Module include some or all of, and variations of, some or all of the scripts set forth below. :

(Example 5)
EXAMPLES OF EMBODIMENTS The following examples describe certain specific embodiments and are not intended to limit the technology.

A1. A system comprising a memory and one or more microprocessors, wherein the one or more microprocessors are configured to perform processing according to instructions in the memory to reduce bias in sequence reads for a sample. and the processing is

(a) generating a relationship between (i) guanine and cytosine (GC) density and (ii) GC density frequency for sequence reads of a test sample, thereby generating a sample GC density relationship;
the sequence reads are of circulating cell-free nucleic acids derived from the test sample;
the sequence reads are mapped against a reference genome;
(b) comparing the sample GC density relationship and the reference GC density relationship, thereby generating a comparison, comprising:
the reference GC density relationship is between (i) the GC density and (ii) the GC density frequency for the reference;

(c) normalizing the sequence read counts for the sample according to the comparisons determined in (b), thereby reducing sequence read bias for the sample.

A1.1. A system comprising a sequencing device and one or more computing devices,
The sequencing device is configured to generate signals corresponding to the nucleotide bases of nucleic acids loaded into the sequencing device, the nucleic acids being circulating cell-free blood derived from the blood of pregnant females giving birth to fetuses. the nucleic acid is a nucleic acid or is a modified variant of a circulating cell-free nucleic acid;
The one or more computing units include memory and one or more processors, the memory including instructions executable by the one or more processors, the instructions executable by the one or more processors comprising:
generate sequence reads from the signal, map the sequence reads against the reference genome,
(a) generating a relationship between (i) guanine and cytosine (GC) density and (ii) GC density frequency for the sequence reads of the test sample, thereby generating a sample GC density relationship;
(b) compare the sample GC density relationship with the reference GC density relationship, thereby producing a comparison (the reference GC density relationship is between (i) the GC density and (ii) the GC density frequency for the reference); is a thing),

(c) a system configured to normalize the sequence read counts for the sample according to the comparison determined in (b), thereby reducing sequence read bias for the sample;

A1.2. The system of embodiment A1 or A1.1, wherein normalizing in (c) comprises providing a normalized count number.

A2. The system as in any one of embodiments A1-A1.2, wherein each of the GC densities is determined by a process that includes using kernel density estimation.

A2.1. The system of any one of embodiments A1-A2, wherein each of the GC densities for the reference GC density relationship and the sample GC density relationship is an indication of local GC content.

A2.2. The system of embodiment A2.1, wherein the local GC content is for polynucleotide segments of 5000 bp or less.

A3. The system of any one of embodiments A1-A2.2, wherein each of the GC densities is determined by a process that includes using a sliding window analysis.

A4. The system of embodiment A3, wherein the window is from about 5 consecutive nucleotides to about 5000 consecutive nucleotides and is slid from about 1 base to about 10 bases simultaneously in the sliding window analysis.

A5. The system of embodiment A3, wherein the window is about 200 contiguous nucleotides and is slid about 1 base at a time in the sliding window analysis.

A6. (b) comprises generating (i) a ratio each comprising a sample GC density-related frequency and a reference GC density-related frequency for each of the GC densities and (ii) a fitted relationship with the GC density; The system of any one of embodiments Al through A5.

A7. The system of embodiment A6, wherein the fitted relationship in (a) is obtained from weighted fitting.

A8. The system of any one of embodiments A1-A7, wherein each of the sequence reads for the sample is displayed in binary and/or text format.

A9. The system of embodiment A8, wherein the binary format for each read of the sequence comprises the chromosome to which the read maps and the chromosomal location to which the read maps.

A10. The system of embodiment A9, wherein the binary format is in a 5-byte format comprising a 1-byte chromosomal ordinal number and a 4-byte chromosomal position.

A11. The system of any one of embodiments A8-A10, wherein the binary format is 50 times smaller than a sequence alignment/map (SAM) format and/or about 13% smaller than a GZip format.

A12. Any one of embodiments A1 through A11, wherein the normalizing in (c) comprises factorizing one or more features other than GC density and normalizing the sequence reads. system.

A13. The system of embodiment A12, wherein the factorization of the one or more features is by a process that includes using a multivariate model.

A14. The system of A13, wherein the processing including the use of multivariate models is performed by a multivariate module.

A14.1. The system of any one of embodiments A12-A14, wherein the sequence read counts are normalized according to the normalization in (c) and the factorization of the one or more features.

A15. (c) followed by generating a probability density estimate for each of the one or more portions comprising the normalized sequence read counts in (c); The system of any one of embodiments A1-A14.1, comprising generating a read density for a portion or segment thereof.

A16. The system of embodiment A15, wherein the probability density estimation is kernel density estimation.

A17. The system of embodiment A15 or A16, comprising generating a read density profile for a genome or segment thereof.

A18. The system of embodiment A17, wherein the read density profile comprises read densities for one or more portions of the genome or segments thereof.

A19. The system of any one of embodiments A15-A18, comprising adjusting each of the lead densities for the one or more portions.

A20. The system of any one of embodiments A15-A19, wherein one or more portions are filtered thereby providing filtered portions.

A21. The system of any one of embodiments A15-A20, wherein one or more portions are weighted, thereby providing weighted portions.

A22. The system of embodiment A21, wherein the one or more portions are weighted by eigenfunctions.

A23. The system of any one of embodiments A1-A22, comprising obtaining sequence reads prior to (a).

A24. The system of embodiment A23, wherein the sequence reads are generated by massively parallel sequencing (MPS).

A25. The system of any one of embodiments A1-A24, comprising obtaining sequence reads mapped to the entire reference genome or a segment of the genome.

A26. The system of embodiment A25, wherein the segment of the genome comprises a chromosome or segment thereof.

A27. The system of embodiment A25 or A26, wherein the read counts of the sequences mapped against the reference genome are normalized prior to (a).

A28. The system of embodiment A27, wherein the read counts of sequences mapped against a reference genome are normalized by GC content, bin-wise normalization, GC LOESS, PERUN, GCRM, or combinations thereof.

A29. The system of any one of embodiments A27 or A28, wherein the read count of the sequence mapped against the reference genome is the raw count.

A30. The system of any one of embodiments A15-A29, wherein each portion of the reference genome comprises approximately equal lengths of contiguous nucleotides.

A31. The system of any one of embodiments A15 or A30, wherein each portion of the reference genome constitutes about 50 kb.

A32. The system of any one of embodiments A15-A31, wherein each portion of the reference genome constitutes about 100 kb.

A33. The system of any one of embodiments A15-A32, wherein each portion of the reference genome comprises a segment of contiguous nucleotides in common with adjacent portions of the reference genome.

A34. The system of any one of embodiments A1-A33, wherein the test sample is obtained from a pregnant female.

A35. The system of any one of embodiments A1-A34, wherein the test sample comprises blood from a pregnant female.

A36. The system of any one of embodiments A1-A35, wherein the test sample comprises plasma from a pregnant female.

A37. The system of any one of Al to A36, wherein the test sample comprises serum from a pregnant female.

A38. The system of any one of embodiments Al through A37, wherein nucleic acids are isolated from the test sample.

A39. The system of any one of embodiments A8-A38, comprising compressing the sequence reads mapped to the reference genome in (a) from a sequence alignment format into a binary format.

A40. The system of embodiment A39, wherein the compression is performed by a compression module.

A41. The system of any one of embodiments A1-A40, wherein the GC densities and GC density frequencies for the test sample sequence reads and for the references are provided by a biased density module.

A42. The system of any one of embodiments A1-A41, wherein the comparison in (b) is generated by a relationship module.

A43. The system of any one of embodiments A1-A42, wherein the normalization in (c) is performed by a bias correction module.

A44. The system of any one of embodiments A15-A43, wherein the lead density is provided by a distribution module.

A45. The system of any one of embodiments A20-A44, wherein the filtered portion is provided by a filtering module.

A46. The system of any one of embodiments A21-A45, wherein the adjusted lead density is provided by a lead density adjustment module.

A46.1. The system as in any one of embodiments A21-A46, wherein the weighted portion is provided by a portion weighting module.

A47. The system of embodiment A46.1 including one or more of a compression module, a bias density module, a relation module, a bias correction module, a distribution module, a filtering module, a read density adjustment module, and a partial weighting module.

A48. The system of any one of embodiments A1-A47, wherein the memory of the system comprises reads of sequences of circulating cell-free nucleic acids from test samples that are mapped against a reference genome.

B1. A system comprising a memory and one or more microprocessors, wherein the one or more microprocessors perform processing to determine the presence or absence of aneuploidy for a sample according to instructions in the memory. It is configured as follows, and the process is

(a) filtering a portion of the reference genome according to the read density distribution, thereby providing a read density profile for the test sample comprising the read density of the filtered portion;
the read density includes reads of sequences of circulating cell-free nucleic acids from test samples from pregnant females;
a read density distribution is determined for the partial read densities for the plurality of samples;
(b) adjusting a read density profile for the test sample according to one or more principal components obtained from a set of known euploid samples by principal component analysis, thereby producing a test sample profile comprising the adjusted read density; providing a step;
(c) comparing the test sample profile to the reference profile, thereby providing a comparison;

(d) determining the presence or absence of chromosomal aneuploidy for the test sample according to the comparison.

B1.1. A system comprising a sequencing device and one or more computing devices,
The sequencing device is configured to generate signals corresponding to the nucleotide bases of nucleic acids loaded into the sequencing device, the nucleic acids being circulating cell-free blood derived from the blood of a pregnant female giving birth to a fetus. the nucleic acid or the nucleic acid loaded into the sequencing device is a modified variant of the circulating cell-free nucleic acid;
The one or more computing units include memory and one or more processors, the memory including instructions executable by the one or more processors, the instructions executable by the one or more processors comprising:
generate sequence reads from the signal, map the sequence reads against the reference genome,
a) filtering a portion of the reference genome according to the read density distribution, thereby providing a read density profile for the test sample comprising the read density of the filtered portion (the read density being derived from the test samples derived from pregnant females); comprising sequences of circular cell-free nucleic acid reads derived from
The lead density distribution is determined for partial lead densities for a plurality of samples),
(b) adjusting a read density profile for the test sample according to one or more principal components obtained from a set of known euploid samples by principal component analysis, thereby producing a test sample profile comprising the adjusted read density; Offer to,
(c) comparing the test sample profile to the reference profile, thereby providing a comparison;

(d) a system configured to determine the presence or absence of a chromosomal aneuploidy for the test sample according to the comparison;

B2. The system of embodiment B1 or B1.1, wherein comparing comprises determining a level of significance.

B3. The system of any one of embodiments B1-B2, wherein determining the level of significance comprises determining a p-value.

B4. The system of any one of embodiments B1-B3, wherein the reference profile comprises a read density profile obtained from a series of known euploid samples.

B5. The system of any one of embodiments B1-B4, wherein the reference profile comprises read densities of the filtered portion.

B6. The system of any one of embodiments B1-B5, wherein the reference profile comprises read densities adjusted according to one or more principal components.

B7. The system of any one of embodiments B2-B6, wherein the level of significance indicates a statistically significant difference between the test sample profile and the reference profile and the presence of a chromosomal aneuploidy is determined. .

B8. The system of any one of embodiments B1-B7, wherein the plurality of samples comprises a series of known euploid samples.

B9. The system of any one of embodiments B1 through B8, wherein the partial lead density for the plurality of samples is the median lead density.

B10. The system of any one of embodiments B1-B9, wherein the read density of the filtered portion for the test sample is the median read density.

B11. The system of any one of embodiments B4-B10, wherein the read density profile for the reference profile comprises a median lead density.

B12. The system of any one of embodiments B4-B11, wherein read densities for the test sample profile, the plurality of samples, and the reference profile are determined according to a process that includes using kernel density estimation.

B13. The system of any one of embodiments B10-B12, wherein the test sample profile is determined according to a median read density for the test samples.

B14. The system of any one of embodiments B11-B13, wherein the reference profile is determined according to a median read density distribution for the reference.

B15. The system of any one of embodiments B1-B14, comprising filtering portions of the reference genome according to a measure of uncertainty about the read density distribution.

B16. The system of embodiment B15, wherein the uncertainty measure is MAD.

B17. The number of sequence read counts mapped to the filtered portion for the test sample is
(I) generating a relationship between (i) local genomic bias estimates and (ii) bias frequencies for sequence reads of a test sample, thereby generating a sample bias relationship,
the sequence reads are of circulating cell-free nucleic acids derived from the test sample;
the sequence reads are mapped against a reference genome;
(II) comparing the sample bias relation and the reference bias relation, thereby generating a comparison, comprising:
the reference bias relationship is between (i) a local genomic bias estimate and (ii) a bias frequency for the reference;

(III) normalizing the sequence read counts for the sample according to the comparison determined in (II), thereby reducing the sequence read bias for the sample; The system of any one of embodiments B1-B16, wherein the system is normalized by a process performed by

B18. The system of embodiment B17, wherein normalizing in (III) includes providing a normalized count number.

B19. The system of embodiment B17 or B18, wherein each of the local genomic bias estimates is determined by a process that includes using kernel density estimation.

B20. The system of any one of embodiments B17-B19, wherein each of the local genomic bias estimates is determined by a process that includes using a sliding window analysis.

B21. The system of embodiment B20, wherein the window is from about 5 consecutive nucleotides to about 5000 consecutive nucleotides and is slid from about 1 base to about 10 bases simultaneously in the sliding window analysis.

B22. The system of embodiment B20, wherein the window is about 200 contiguous nucleotides and is slid about 1 base at a time in the sliding window analysis.

B23. (II) is a ratio comprising (i) a sample-biased related frequency and a reference-biased related frequency, each for each of the local genomic bias estimates; and (ii) the local genomic bias estimate; The system of any one of embodiments B17-B22, comprising generating the matched relationship of .

B24. The system of embodiment B23, wherein the fitted relationship in (I) is obtained from weighted fitting.

B25. The system of any one of embodiments B17 through B24, wherein each of the sequence reads for the sample is displayed in binary format.

B26. The system of embodiment B25, wherein the binary format for each read of the sequence comprises the chromosome to which the read maps and the chromosomal location to which the read maps.

B27. The system of embodiment B26, wherein the binary format is in a 5-byte format comprising a 1-byte chromosomal ordinal number and a 4-byte chromosomal position.

B28. The system of any one of embodiments B25-B27, wherein the binary format is 50 times smaller than a sequence alignment/map (SAM) format and/or about 13% smaller than a GZip format.

B29. any one of embodiments B17 through B28, wherein the normalizing in (III) comprises factorizing one or more features other than bias and normalizing the sequence read counts System as described.

B30. The system of embodiment B29, wherein the factorization of the one or more features is by a process that includes using a multivariate model.

B31. The system of embodiment B30, wherein the processing including the use of multivariate models is performed by a multivariate module.

B32. The system of any one of embodiments B29-B31, wherein the sequence read counts are normalized according to the normalization in (III) and the factorization of the one or more features.

B33. (III) followed by generating a probability density estimate for each of the one or more portions comprising the sequence read counts normalized in (III). The system of any one of embodiments B17-B32, comprising generating a read density for a portion or segment thereof.

B34. The system of embodiment B33, wherein the probability density estimation is kernel density estimation.

B35. The system of embodiment B33 or B34, comprising generating a read density profile for a genome or segment thereof.

B36. The system of embodiment B35, wherein the read density profile comprises read densities for one or more portions of the genome or segments thereof.

B37. The system of any one of embodiments B33-B36, comprising adjusting each of the lead densities for the one or more portions.

B38. The system of any one of embodiments B33-B37, wherein one or more portions are filtered thereby providing filtered portions.

B39. The system of any one of embodiments B33-B38, wherein one or more portions are weighted, thereby providing weighted portions.

B40. The system of embodiment B39, wherein the one or more portions are weighted by eigenfunctions.

B41. The system of any one of embodiments B17-B40, wherein the local genomic bias estimate is the local GC density and the bias frequency is the GC bias frequency.

B42. The number of sequence read counts mapped to the filtered portion for the test sample is

(1) generating a fitted relationship between (i) guanine and cytosine (GC) densities and (ii) GC density frequencies for sequence reads of a test sample, thereby generating a sample GC density relationship. the sequence reads are mapped against a reference genome;
(2) comparing the sample GC density relationship and the reference GC density relationship, thereby generating a comparison, comprising:
that the reference GC density relationship is between (i) the GC density and (ii) the GC density frequency for the reference;

(3) normalizing the sequence read counts for the sample according to the comparison determined in (b), thereby reducing the sequence read bias for the sample; The system of any one of embodiments B1-B16, wherein the system is normalized by processing.

B43. The system of embodiment B42, wherein the normalizing in (3) includes providing a normalized count number.

B44. The system of embodiment B42 or B43, wherein each of the GC densities is determined by a process that includes using kernel density estimation.

B44.1. The system of any one of embodiments B42-B44, wherein each of the GC densities for the reference GC density relationship and the sample GC density relationship is an indication of local GC content.

B44.2. The system of embodiment B44.1, wherein the local GC content is for polynucleotide segments of 5000 bp or less.

B45. The system of any one of embodiments B42-B44.2, wherein each of the GC densities is determined by a process that includes using a sliding window analysis.

B46. The system of embodiment B45, wherein the window is from about 5 contiguous nucleotides to about 5000 contiguous nucleotides and is slid from about 1 base to about 10 bases simultaneously in a sliding window analysis.

B47. The system of embodiment B46, wherein the window is about 200 contiguous nucleotides and is slid about 1 base at a time in the sliding window analysis.

B48. (2) comprises generating (i) a ratio each comprising a sample GC density-related frequency and a reference GC density-related frequency for each of the GC densities and (ii) a fitted relationship with the GC density; The system of any one of embodiments B42-B47.

B49. The system of embodiment B48, wherein the fitted relationship in (1) is obtained from weighted fitting.

B50. The system of any one of embodiments B42-B49, wherein each of the sequence reads for the sample is displayed in a binary format.

B51. The system of embodiment B50, wherein the binary format for each read of the sequence comprises the chromosome to which the read maps and the chromosomal location to which the read maps.

B52. The system of embodiment B51, wherein the binary format is in a 5-byte format comprising a 1-byte chromosomal ordinal number and a 4-byte chromosomal position.

B53. The system of any one of embodiments B50-B52, wherein the binary format is 50 times smaller than a sequence alignment/map (SAM) format and/or about 13% smaller than a GZip format.

B54. Any one of embodiments B42-B53, wherein the normalizing in (c) comprises factorizing one or more features other than GC density and normalizing the sequence reads. system.

B55. The system of embodiment B54, wherein the factorization of the one or more features is by a process that includes using a multivariate model.

B56. The system of embodiment B55, wherein the processing including the use of multivariate models is performed by a multivariate module.

B57. The system of any one of embodiments B42-B56, wherein the filtered portion of the test sample is weighted.

B58. The system of embodiment B57, wherein the filtered portion of the test sample is weighted by a process that includes an eigenfunction.

B59. The system of any one of embodiments B1-B58, comprising obtaining sequence reads prior to (a).

B60. The system of embodiment B59, wherein the sequence reads are generated by massively parallel sequencing (MPS).

B61. The system of any one of embodiments B1-B60, comprising obtaining sequence reads mapped to the entire reference genome or a segment of the genome.

B62. The system of embodiment B61, wherein the segment of the genome comprises a chromosome or segment thereof.

B63. The system of embodiment B61 or B62, wherein the read counts of the sequences mapped against the reference genome are normalized prior to (1).

B64. The system of embodiment B63, wherein the read counts of sequences mapped against a reference genome are normalized by GC content, bin-wise normalization, GC LOESS, PERUN, GCRM, or a combination thereof.

B65. The system of embodiment B61 or B62, wherein the read count of the sequence mapped against the reference genome is the raw count.

B66. The system of any one of embodiments B1-B65, wherein each portion of the reference genome comprises approximately equal lengths of contiguous nucleotides.

B67. The system of any one of embodiments B1-B66, wherein each portion of the reference genome constitutes about 50 kb.

B68. The system of any one of embodiments B1-B67, wherein each portion of the reference genome constitutes about 100 kb.

B69. The system of any one of embodiments B1-B68, wherein each portion of the reference genome comprises a segment of contiguous nucleotides in common with adjacent portions of the reference genome.

B70. The system of any one of embodiments B1-B69, wherein the test sample comprises blood from a pregnant female.

B71. The system of any one of embodiments B1-B70, wherein the test sample comprises plasma from a pregnant female.

B72. The system of any one of embodiments B1-B71, wherein the test sample comprises serum from a pregnant female.

B73. The system of any one of embodiments B1 through B72, wherein nucleic acids are isolated from the test sample.

B74. The system of any one of embodiments B50-B73, comprising compressing the sequence reads mapped to the reference genome in (1) from a sequence alignment format into a binary format.

B75. The system of embodiment B74, wherein the compression is performed by a compression module.

B76. The system of any one of embodiments B42-B75, wherein the GC densities and GC density frequencies for the test sample sequence reads and for the references are provided by a biased density module.

B77. The system of any one of embodiments B42-B76, wherein the comparison in (2) is generated by a relationship module.

B78. The system of any one of embodiments B44-B77, wherein the normalization in (3) is performed by a bias correction module.

B79. The system of any one of embodiments B1-B78, wherein the lead density is provided by a distribution module.

B80. The system according to any one of embodiments B1-B79, wherein the filtered portion is provided by a filtering module.

B81. The system of any one of embodiments B57-B80, wherein the filtered portions for the test sample are weighted by a portion weighting module.

B81.1. The system of any one of embodiments B57-B81, wherein lead density is adjusted by a lead density adjustment module.

B82. The apparatus of embodiment B81.1, wherein the apparatus includes one or more of a compression module, a bias density module, a relationship module, a bias correction module, a distribution module, a filtering module, a read density adjustment module, and a partial weighting module. system.

B83. The system of any one of embodiments B1-B82, wherein the test sample profile comprises a profile of a chromosome or segment thereof.

B84. The system of any one of embodiments B1-B83, wherein the reference profile comprises a profile of a chromosome or segment thereof.

B85. The system of any one of embodiments B1 through B84, wherein the determination in (d) is provided with a specificity equal to or greater than 90% and a sensitivity equal to or greater than 90%.

B86. The system of any one of embodiments B1-B85, wherein the aneuploidy is a trisomy.

B87. The system of embodiment B86, wherein the trisomy is trisomy 21, trisomy 18, or trisomy 13.

B88. The system of any one of embodiments B17 through B87, wherein the memory of the system comprises reads of sequences of circulating cell-free nucleic acids from test samples that are mapped against a reference genome.

C1. The system of any one of embodiments A1-A48 and B1-B88 incorporated into one or more machines.

C2. The system of embodiment C1 integrated into one machine.

C3. The system of embodiment C1 or C2, comprising a machine configured to sequence nucleic acids and generate sequence reads.

D1. A method for reducing sequence read bias for a sample, comprising:
(a) using a microprocessor to generate a relationship between (i) guanine and cytosine (GC) density and (ii) GC density frequency for the sequence reads of the test sample, thereby generating a sample GC density relationship; a step of
the sequence reads are of circulating cell-free nucleic acids derived from the test sample;
the sequence reads are mapped against a reference genome;
(b) comparing the sample GC density relationship and the reference GC density relationship, thereby generating a comparison, comprising:
the reference GC density relationship is between (i) the GC density and (ii) the GC density frequency for the reference;

(c) normalizing the sequence read counts for the sample according to the comparison determined in (b), thereby reducing sequence read bias for the sample.

D1.1. A method for reducing sequence read bias for a sample, comprising:
Loading a circulating cell-free nucleic acid derived from the blood of a fetus-bearing pregnant female into a sequencing device, or loading a modified variant of the nucleic acid into a sequencing device, wherein the sequencing device generating a signal corresponding to the nucleotide base of
generating sequence reads from the nucleic acid signal, optionally after transfer of the signal to the system, by a system comprising one or more computing units, wherein the one or more computing units in the system: including memory and one or more processors;
One computing unit or combination of computing units in the system
mapping the sequence reads against a reference genome;
(a) generating a relationship between (i) guanine and cytosine (GC) density and (ii) GC density frequency for sequence reads of test samples, thereby generating a sample GC density relationship (sequence reads are of circulating cell-free nucleic acids derived from a test sample;
Sequence reads are mapped against a reference genome),
(b) comparing the sample GC density relationship with the reference GC density relationship, thereby generating a comparison (the reference GC density relationship being between (i) the GC density and (ii) the GC density frequency for the reference; is),

(c) normalizing the sequence read counts for the sample according to the comparison determined in (b), thereby reducing sequence read bias for the sample. .

D1.2. The method of embodiment D1 or D1.1, wherein normalizing in (c) comprises providing a normalized count number.

D2. The method as in any one of embodiments D1-D1.2, wherein each of the GC densities is determined by a process that includes using kernel density estimation.

D2.1. The method of any one of embodiments D1-D2, wherein each of the GC densities for the reference GC density relationship and the sample GC density relationship is an indication of local GC content.

D2.2. The method of embodiment D2.1, wherein the local GC content is for polynucleotide segments of 5000 bp or less.

D3. The method of any one of embodiments D1-D2.2, wherein each of the GC densities is determined by a process that includes using a sliding window analysis.

D4. The method of embodiment D3, wherein the window is from about 5 contiguous nucleotides to about 5000 contiguous nucleotides and is slid from about 1 base to about 10 bases simultaneously in the sliding window analysis.

D5. The method of embodiment D3, wherein the window is about 200 contiguous nucleotides and is slid about 1 base at a time in the sliding window analysis.

D6. (b) comprises generating (i) a ratio each comprising a sample GC density-related frequency and a reference GC density-related frequency for each of the GC densities and (ii) a fitted relationship with the GC density; The method of any one of embodiments D1 through D5.

D7. The method of embodiment D6, wherein the fitted relationship in (a) is obtained from weighted fitting.

D8. The method of any one of embodiments D1-D7, wherein each of the sequence reads for the sample is displayed in binary format.

D9. The method of embodiment D8, wherein the binary format for each read of the sequence comprises the chromosome to which the read maps and the chromosomal location to which the read maps.

D10. The method of embodiment D9, wherein the binary format is in a 5-byte format comprising a 1-byte chromosomal ordinal number and a 4-byte chromosomal position.

D11. The method of any one of embodiments D8 through D10, wherein the binary format is 50 times smaller than a sequence alignment/map (SAM) format and/or about 13% smaller than a GZip format.

D12. Any one of embodiments D1 through D11, wherein the normalizing in (c) comprises factorizing one or more features other than GC density and normalizing sequence read counts The method described in .

D13. The method of embodiment D12, wherein the factorization of the one or more features is by a process that includes using a multivariate model.

D14. The method of D13, wherein the processing including the use of multivariate models is performed by a multivariate module.

D14.1. The method of any one of embodiments D12-D14, wherein the sequence read counts are normalized according to the normalization in (c) and the factorization of the one or more features.

D15. (c) followed by generating a probability density estimate for each of the one or more portions comprising the normalized sequence read counts in (c); The method of any one of embodiments D1 through D14.1, comprising generating a read density for a portion or segment thereof.

D16. The method of embodiment D15, wherein the probability density estimation is kernel density estimation.

D17. The method of embodiment D15 or D16, comprising generating a read density profile for a genome or segment thereof.

D18. The method of embodiment D17, wherein the read density profile comprises read densities for one or more portions of the genome or segments thereof.

D19. The method of any one of embodiments D15-D18, comprising adjusting each of the lead densities for the one or more portions.

D20. The method of any one of embodiments D15-D19, wherein one or more portions are filtered, thereby providing filtered portions.

D21. The method of any one of embodiments D15-D20, wherein the one or more portions are weighted, thereby providing weighted portions.

D22. The method of embodiment D21, wherein the one or more portions are weighted by eigenfunctions.

D23. The method of any one of embodiments D1 through D22, prior to (a), comprising obtaining sequence reads.

D24. The method of embodiment D23, wherein the sequence reads are generated by massively parallel sequencing (MPS).

D25. The method of any one of embodiments D1-D24, comprising obtaining sequence reads mapped to the entire reference genome or a segment of the genome.

D26. The method of embodiment D25, wherein the segment of the genome comprises a chromosome or segment thereof.

D27. The method of embodiment D25 or D26, wherein the read counts of the sequences mapped against the reference genome are normalized prior to (a).

D28. The method of embodiment D27, wherein the read counts of the sequences mapped against the reference genome are normalized by GC content, bin-wise normalization, GC LOESS, PERUN, GCRM, or combinations thereof.

D29. The method of any one of embodiments D27 or D28, wherein the read counts of the sequences mapped against the reference genome are raw counts.

D30. The method of any one of embodiments D15-D29, wherein each portion of the reference genome comprises approximately equal lengths of contiguous nucleotides.

D31. The method of any one of embodiments D15 or D30, wherein each portion of the reference genome constitutes about 50 kb.

D32. The method of any one of embodiments D15-D31, wherein each portion of the reference genome constitutes about 100 kb.

D33. The method of any one of embodiments D15-D32, wherein each portion of the reference genome comprises a segment of contiguous nucleotides in common with adjacent portions of the reference genome.

D34. The method of any one of embodiments D1 through D33, wherein the test sample is obtained from a pregnant female.

D35. The method of any one of embodiments D1-D34, wherein the test sample comprises blood from a pregnant female.

D36. The method of any one of embodiments D1-D35, wherein the test sample comprises plasma from a pregnant female.

D37. The method of any one of embodiments D1-D36, wherein the test sample comprises serum from a pregnant female.

D38. The method of any one of embodiments D1 through D37, wherein the nucleic acid is isolated from the test sample.

D39. The method of any one of embodiments D8-D38, comprising compressing the sequence reads mapped to the reference genome in (a) from a sequence alignment format into a binary format.

D40. The method of embodiment D39, wherein the compression is performed by a compression module.

D41. The method of any one of embodiments D1-D40, wherein the GC densities and GC density frequencies for the test sample sequence reads and for the references are provided by a biased density module.

D42. The method of any one of embodiments D1-D41, wherein the comparison in (b) is generated by a relationship module.

D43. The method of any one of embodiments D1-D42, wherein the normalization in (c) is performed by a bias correction module.

D44. The method of any one of embodiments D15-D43, wherein the lead density is provided by a distribution module.

D45. The method of any one of embodiments D20-D44, wherein the filtered portion is provided by a filtering module.

D46. The method of any one of embodiments D21-D45, wherein the weighted portion is provided by a portion weighting module.

D46.1. The method of any one of embodiments D21-D46, wherein the lead density is adjusted by a lead density adjustment module.

D47. The method of embodiment D46.1 including one or more of a compression module, a bias density module, a relation module, a bias correction module, a distribution module, a filtering module, a read density adjustment module, and a partial weighting module.

E0. A method for determining the presence or absence of aneuploidy for a sample, comprising:
(a) filtering a portion of the reference genome according to the read density distribution, thereby providing a read density profile for the test sample comprising the read density of the filtered portion;
the read density includes reads of sequences of circulating cell-free nucleic acids from test samples from pregnant females;
a read density distribution is determined for the partial read densities for the plurality of samples;
(b) adjusting a read density profile for the test sample according to one or more principal components obtained from a set of known euploid samples by principal component analysis, thereby producing a test sample profile comprising the adjusted read density; providing a step;
(c) comparing the test sample profile to the reference profile, thereby providing a comparison;

(d) determining the presence or absence of chromosomal aneuploidy for the test sample according to the comparison.

E0.1. A method for determining the presence or absence of aneuploidy for a sample, comprising:
(a) filtering portions of the chromosomes in the reference genome according to the read density distribution, thereby providing a read density profile for the test sample comprising the read densities of the filtered portions;
the read density comprises reads of sequences of circulating cell-free nucleic acids from test samples from pregnant females;
a lead density distribution is determined for the partial lead densities for the plurality of samples;
(b) adjusting a chromosomal read density profile for the test sample according to one or more principal components obtained from a set of known euploid samples by principal component analysis, whereby the test sample comprising the adjusted read density; providing a chromosomal profile;
(c) comparing the test sample chromosomal profile to the reference profile, thereby providing a comparison;

(d) determining the presence or absence of chromosomal aneuploidy for the test sample according to the comparison.

E1. A method for determining the presence or absence of aneuploidy for a sample, comprising:
(a) filtering a portion of the reference genome according to the read density distribution, thereby providing a read density profile for the test sample comprising the read density of the filtered portion;
the read density comprises reads of sequences of circulating cell-free nucleic acids from test samples from pregnant females;
a lead density distribution is determined for the partial lead densities for the plurality of samples;
(b) using a microprocessor to adjust the read density profile for the test sample according to one or more principal components obtained from a series of known euploid samples by principal component analysis, thereby adjusting the read density profile; providing a test sample profile including density;
(c) comparing the test sample profile to the reference profile, thereby providing a comparison;

(d) determining the presence or absence of chromosomal aneuploidy for the test sample according to the comparison.

E1.1. A method for determining the presence or absence of aneuploidy for a sample, comprising:
Loading a circulating cell-free nucleic acid derived from the blood of a fetus-bearing pregnant female into a sequencing device, or loading a modified variant of the nucleic acid into a sequencing device, wherein the sequencing device generating a signal corresponding to the nucleotide base of
generating sequence reads from the nucleic acid signal, optionally after transfer of the signal to the system, by a system comprising one or more computing units, wherein the one or more computing units in the system: including memory and one or more processors;
one computing unit or combination of computing units in the system maps the sequence reads to a reference genome;
(a) filtering a portion of the reference genome according to the read density distribution, thereby providing a read density profile for the test sample comprising the read densities of the filtered portion (read densities are derived from test samples derived from pregnant females); comprising a sequence read of circulating cell-free nucleic acid derived from the sample;
The lead density distribution is determined for partial lead densities for a plurality of samples),
(b) using a microprocessor to adjust the read density profile for the test sample according to one or more principal components obtained from a series of known euploid samples by principal component analysis, thereby adjusting the read density profile; providing a test sample profile including density,
(c) comparing the test sample profile to the reference profile, thereby providing a comparison;

(d) a method configured to determine the presence or absence of a chromosomal aneuploidy for the test sample according to the comparison.

E1.2. A method for reducing sequence read bias for a sample, comprising:
loading a circulating cell-free nucleic acid derived from the blood of a fetus-bearing pregnant female into a sequencing device, or loading a modified variant of the nucleic acid into a sequencing device, wherein the sequencing device generating a signal corresponding to the nucleotide base of
generating sequence reads from the nucleic acid signal, optionally after transfer of the signal to the system, by a system comprising one or more computing units, wherein the one or more computing units in the system: including memory and one or more processors;
one computing unit or combination of computing units in the system maps the sequence reads to the reference genome;
(a) filtering a portion of the reference genome according to the read density distribution, thereby providing a read density profile for the test sample comprising the read densities of the filtered portion (read densities are derived from test samples derived from pregnant females); comprising a sequence read of circulating cell-free nucleic acid derived from the sample;
The lead density distribution is determined for partial lead densities for a plurality of samples),
(b) using a microprocessor to adjust a read density profile for the test sample according to one or more principal components obtained from a set of known euploid samples by principal component analysis, thereby adjusting the read density profile; providing a test sample profile including density,
(c) comparing the test sample profile to the reference profile, thereby providing a comparison;

(d) a method configured to determine the presence or absence of a chromosomal aneuploidy for the test sample according to the comparison.

E1.3. The method of any one of embodiments E0 through E1.2, wherein the read density profile is adjusted in (b) by 1-10 principal components.

E1.4. The method of any one of embodiments E0 through E1.3, wherein the read density profile is adjusted in (b) by five principal components.

E1.5. One or more principal components adjusted for one or more features in the read density profile, the features correlated to fetal sex, sequence bias, fetal fraction, DNase I sensitivity bias , entropy, repetitive sequence bias, chromatin structure bias, polymerase error rate bias, batch sequence bias, inverted repeat bias, PCR amplification bias, and hidden copy number variation, embodiment EO E1.4.

E1.6. The method of embodiment E1.5, wherein the sequence bias comprises a guanine and cytosine (GC) bias.

E2. The method of any one of embodiments E0 through E1.6, wherein comparing comprises determining a level of significance.

E3. The method of any one of embodiments E0-E2, wherein determining the level of significance comprises determining a p-value.

E4. The method of any one of embodiments E0-E3, wherein the reference profile comprises a read density profile obtained from a series of known euploid samples.

E5. The method of any one of embodiments E0-E4, wherein the reference profile comprises read densities of the filtered portion.

E6. The method of any one of embodiments E0-E5, wherein the reference profile comprises read densities adjusted according to one or more principal components.

E7. The method of any one of embodiments E2-E6, wherein the level of significance indicates a statistically significant difference between the test sample profile and the reference profile and the presence of a chromosomal aneuploidy is determined. .

E8. The method of any one of embodiments E1 through E7, wherein the plurality of samples comprises a series of known euploid samples.

E9. The method of any one of embodiments E0 through E8, wherein the partial lead density for the plurality of samples is the median lead density.

E10. The method of any one of embodiments E0-E9, wherein the read density of the filtered portion for the test sample is the median read density.

E11. The method of any one of embodiments E4-E10, wherein the read density profile for the reference profile comprises a median lead density.

E12. The method of any one of embodiments E4-E11, wherein read densities for the test sample profile, the plurality of samples, and the reference profile are determined according to a process that includes using kernel density estimation.

E13. The method of any one of embodiments E10-E12, wherein the test sample profile is determined according to the median read density for the test samples.

E14. The method of any one of embodiments E11-E13, wherein the reference profile is determined according to a median read density distribution for the reference.

E15. The method of any one of embodiments E0-E14, comprising filtering portions of the reference genome according to an uncertainty measure for the read density distribution.

E16. The method of embodiment E15, wherein the uncertainty measure is MAD.

E16.1. The method of any one of embodiments E0-E16, wherein the test sample profile is a representation of chromosomal dose for the test sample.

E16.2. The method of embodiment E16.1, comprising comparing the chromosome dose for the test sample profile to the chromosome dose for the reference profile, thereby generating a chromosome dose comparison.

E16.3. The method of embodiment E16.2, wherein determining the presence or absence of chromosomal aneuploidy for the test sample is according to a comparison of chromosome dose.

E16.4. A determination of the presence or absence of a chromosomal aneuploidy for a test sample results in 1 copy of the chromosome, 2 copies of the chromosome, 3 copies of the chromosome, 4 copies of the chromosome, 5 copies of the chromosome, 1 copy of the chromosome The method of any one of embodiments E0 through E16.3, comprising identifying the presence or absence of deletions of one or more segments or insertions of one or more segments of the chromosome.

E17. The number of sequence read counts mapped to the filtered portion for the test sample is
(I) generating a relationship between (i) local genomic bias estimates and (ii) bias frequencies for sequence reads of a test sample, thereby generating a sample bias relationship,
the sequence reads are of circulating cell-free nucleic acids derived from the test sample;
the sequence reads are mapped against a reference genome;
(II) comparing the sample bias relationship and the reference bias relationship, thereby generating a comparison, comprising:
the reference bias relationship is between (i) a local genomic bias estimate and (ii) a bias frequency for the reference;

(III) normalizing the sequence read counts for the sample according to the comparison determined in (II), thereby reducing the sequence read bias for the sample; The method of any one of embodiments E0 through E16.4, wherein the normalization is performed by the processing performed by

E18. The method of embodiment E17, wherein the normalizing in (III) comprises providing normalized counts.

E19. The method of embodiment E17 or E18, wherein each of the local genomic bias estimates is determined by a process comprising using kernel density estimation.

E19.1. The method of any one of embodiments E17-E19, wherein each of the local genomic bias estimates for the reference bias relationship and the sample bias relationship is an indication of local bias content.

E19.2. The method of embodiment E19.1, wherein the local bias content is for polynucleotide segments of 5000 bp or less.

E20. The method of any one of embodiments E17-E19.2, wherein each of the local genomic bias estimates is determined by a process that includes using a sliding window analysis.

E21. The method of embodiment E20, wherein the window is from about 5 contiguous nucleotides to about 5000 contiguous nucleotides and is slid from about 1 base to about 10 bases simultaneously in a sliding window analysis.

E22. The method of embodiment E20, wherein the window is about 200 contiguous nucleotides and is slid about 1 base at a time in the sliding window analysis.

E23. (II) is a ratio comprising (i) a sample-biased related frequency and a reference-biased related frequency, each for each of the local genomic bias estimates; and (ii) the local genomic bias estimate; The method as in any one of embodiments E17-E22, comprising generating the fitted relationship of .

E24. The method of embodiment E23, wherein the fitted relationship in (I) is obtained from weighted fitting.

E25. The method of any one of embodiments E17-E24, wherein each of the sequence reads for the sample is displayed in binary format.

E26. The method of embodiment E25, wherein the binary format for each read of the sequence comprises the chromosome to which the read maps and the chromosomal location to which the read maps.

E27. The method of embodiment E26, wherein the binary format is in a 5-byte format comprising a 1-byte chromosomal ordinal number and a 4-byte chromosomal position.

E28. The method of any one of embodiments E25-E27, wherein the binary format is 50 times smaller than a sequence alignment/map (SAM) format and/or about 13% smaller than a GZip format.

E29. any one of embodiments E17 through E28, wherein the normalizing in (III) comprises factorizing one or more features other than bias and normalizing the sequence read counts described method.

E30. The method of embodiment E29, wherein the factorization of the one or more features is by a process that includes using a multivariate model.

E31. The method of E30, wherein the processing including the use of multivariate models is performed by a multivariate module.

E32. The method of any one of embodiments E29-E31, wherein the sequence read counts are normalized according to the normalization in (III) and the factorization of the one or more features.

E33. (III), followed by generating a probability density estimate for each of the one or more portions comprising one or more of the sequence read counts normalized in (III), of the genome. The method of any one of embodiments E17-E32, comprising generating read densities for one or more portions or segments thereof.

E34. The method of embodiment E33, wherein the probability density estimation is kernel density estimation.

E35. The method of embodiment E33 or E34, comprising generating a read density profile for a genome or segment thereof.

E36. The method of embodiment E35, wherein the read density profile comprises read densities for one or more portions of the genome or segments thereof.

E37. The method of any one of embodiments E33-E36, comprising adjusting each of the lead densities for the one or more portions.

E38. The method of any one of embodiments E33-E37, wherein one or more portions are filtered thereby providing filtered portions.

E39. The method of any one of embodiments E33-E38, wherein the one or more portions are weighted, thereby providing weighted portions.

E40. The method of embodiment E39, wherein the one or more portions are weighted by eigenfunctions.

E41. The method of any one of embodiments E17-E40, wherein the local genomic bias estimate is the local GC density and the bias frequency is the GC bias frequency.

E42. The number of sequence read counts mapped to the filtered portion for the test sample is

(1) generating a fitted relationship between (i) guanine and cytosine (GC) density and (ii) GC density frequency for the sequence reads of the test sample, thereby generating a sample GC density relationship. the sequence reads are mapped against a reference genome;
(2) comparing the sample GC density relationship and the reference GC density relationship, thereby generating a comparison, comprising:
the reference GC density relationship is between (i) the GC density and (ii) the GC density frequency for the reference;

(3) normalizing the sequence read counts for the sample according to the comparison determined in (b), thereby reducing the sequence read bias for the sample; The method as in any one of embodiments E0-E16, normalized by processing.

E43. The method of embodiment E42, wherein the normalizing in (3) includes providing a normalized count number.

E44. The method of embodiment E42 or E43, wherein each of the GC densities is determined by a process that includes using kernel density estimation.

E45. The method of any one of embodiments E42-E44, wherein each of the GC densities is determined by a process that includes using a sliding window analysis.

E46. The method of embodiment E45, wherein the window is from about 5 contiguous nucleotides to about 5000 contiguous nucleotides and is slid from about 1 base to about 10 bases simultaneously in a sliding window analysis.

E47. The method of embodiment E46, wherein the window is about 200 contiguous nucleotides and is slid about 1 base at a time in the sliding window analysis.

E48. (2) comprises generating (i) a ratio each comprising a sample GC density-related frequency and a reference GC density-related frequency for each of the GC densities and (ii) a fitted relationship with the GC density; The method of any one of embodiments E42-E47.

E49. The method of embodiment E48, wherein the fitted relationship in (1) is obtained from weighted fitting.

E50. The method of any one of embodiments E42-E49, wherein each of the sequence reads for the sample is displayed in binary format.

E51. The method of embodiment E50, wherein the binary format for each read of the sequence comprises the chromosome to which the read maps and the chromosomal location to which the read maps.

E52. The method of embodiment E51, wherein the binary format is in a 5-byte format comprising a 1-byte chromosomal ordinal number and a 4-byte chromosomal position.

E53. The method of any one of embodiments E50-E52, wherein the binary format is 50 times smaller than a sequence alignment/map (SAM) format and/or about 13% smaller than a GZip format.

E54. Any one of embodiments E42-E53, wherein normalizing in (c) comprises factorizing one or more features other than GC density and normalizing sequence reads. Method.

E55. The method of embodiment E54, wherein the factorization of the one or more features is by a process that includes using a multivariate model.

E56. The method of embodiment E55, wherein the processing including using the multivariate model is performed by a multivariate module.

E57. The method of any one of embodiments E42-E56, wherein the filtered portions for the test sample are weighted.

E58. The method of embodiment E57, wherein the filtered portion for the test sample is weighted by a process that includes an eigenfunction.

E59. The method of any one of embodiments E0-E58, prior to (a), comprising obtaining sequence reads.

E60. The method of embodiment E59, wherein the sequence reads are generated by massively parallel sequencing (MPS).

E61. The method of any one of embodiments E0-E60, comprising obtaining sequence reads mapped to the entire reference genome or a segment of the genome.

E62. The method of embodiment E61, wherein the segment of the genome comprises a chromosome or segment thereof.

E63. The method of embodiment E61 or E62, wherein the read counts of the sequences mapped against the reference genome are normalized prior to (1).

E64. The method of embodiment E63, wherein the read counts of the sequences mapped against the reference genome are normalized by GC content, bin-wise normalization, GC LOESS, PERUN, GCRM, or combinations thereof.

E65. The method of embodiment E61 or E62, wherein the read counts of the sequences mapped against the reference genome are raw counts.

E66. The method of any one of embodiments E0-E65, wherein each portion of the reference genome comprises approximately equal lengths of contiguous nucleotides.

E67. The method of any one of embodiments E0-E66, wherein each portion of the reference genome constitutes about 50 kb.

E68. The method of any one of embodiments E0-E67, wherein each portion of the reference genome constitutes about 100 kb.

E69. The method of any one of embodiments E0-E68, wherein each portion of the reference genome comprises a segment of contiguous nucleotides in common with adjacent portions of the reference genome.

E70. The method of any one of embodiments E0-E69, wherein the test sample comprises blood from a pregnant female.

E71. The method of any one of embodiments E0 through E70, wherein the test sample comprises plasma from a pregnant female.

E72. The method of any one of embodiments E0-E71, wherein the test sample comprises serum from a pregnant female.

E73. The method of any one of embodiments E0 through E72, wherein the nucleic acid is isolated from the test sample.

E74. The method of any one of embodiments E50-E73, comprising compressing the sequence reads mapped to the reference genome in (1) from a sequence alignment format to a binary format.

E75. The method of embodiment E74, wherein the compression is performed by a compression module.

E76. The method of any one of embodiments E42-E75, wherein the GC densities and GC density frequencies for the test sample sequence reads and for the references are provided by a biased density module.

E77. The method of any one of embodiments E42 through E76, wherein the comparison in (2) is generated by a relationship module.

E78. The method of any one of embodiments E44-E77, wherein the normalization in (3) is performed by a bias correction module.

E79. The method of any one of embodiments E0-E78, wherein the lead density is provided by a distribution module.

E80. The method of any one of embodiments E0-E79, wherein the filtered portion is provided by a filtering module.

E81. The method of any one of embodiments E57-E80, wherein the filtered portions for the test sample are weighted by a portion weighting module.

E81.1. The method of any one of embodiments E57-E81, wherein the lead density is adjusted by a lead density adjustment module.

E82. The method of embodiment E81.1, wherein the apparatus includes one or more of a compression module, a bias density module, a relationship module, a bias correction module, a distribution module, a filtering module, a read density adjustment module, and a partial weighting module. Method.

E83. The method of any one of embodiments E0-E82, wherein the test sample profile comprises a profile of a chromosome or segment thereof.

E84. The method of any one of embodiments E0-E83, wherein the reference profile comprises a profile of a chromosome or segment thereof.

E85. The method of any one of embodiments E0 through E84, wherein the determination in (d) is provided with a specificity equal to or greater than 90% and a sensitivity equal to or greater than 90%.

E86. The method of any one of embodiments E0-E85, wherein the aneuploidy is a trisomy.

E87. The method of embodiment E86, wherein the trisomy is trisomy 21, trisomy 18, or trisomy 13.

F1. A non-transitory computer-readable storage medium containing a self-stored executable program, the program stored in a microprocessor,
(a) generating a relationship between (i) guanine and cytosine (GC) density and (ii) GC density frequency for sequence reads of a test sample, thereby generating a sample GC density relationship,
the sequence reads are of circulating cell-free nucleic acid derived from the test sample;
the sequence reads are mapped against a reference genome;
(b) comparing a sample GC density relationship with a reference GC density relationship, thereby generating a comparison, wherein the reference GC density relationship is (i) the GC density and (ii) the GC density frequency for the reference; and that it is between

(c) normalizing the sequence read counts for the sample according to the comparison determined in (b), thereby reducing the sequence read bias for the sample; computer readable storage media.

F1.1. The storage medium of embodiment F1, wherein the normalizing in (c) includes providing a normalized count of reads.

F2. The storage medium of embodiment F1 or F1.1, wherein each of the GC densities is determined by a process that includes using kernel density estimation.

F2.1. The storage medium of any one of embodiments F1-F2, wherein each of the GC densities for the reference GC density relationship and the sample GC density relationship is an indication of local GC content.

F2.2. The storage medium of embodiment F2.1, wherein the local GC content is for polynucleotide segments of 5000 bp or less.

F3. The storage medium of any one of embodiments F1-F2.2, wherein each of the GC densities is determined by a process that includes using a sliding window analysis.

F4. The storage medium of embodiment F3, wherein the window is from about 5 contiguous nucleotides to about 5000 contiguous nucleotides and is slid from about 1 base to about 10 bases simultaneously in a sliding window analysis.

F5. The storage medium of embodiment F3, wherein the window is about 200 contiguous nucleotides and is slid about 1 base at a time in the sliding window analysis.

F6. (b) comprises generating (i) a ratio each comprising a sample GC density-related frequency and a reference GC density-related frequency for each of the GC densities and (ii) a fitted relationship with the GC density; The storage medium of any one of embodiments F1 through F5.

F7. The storage medium of embodiment F6, wherein the fitted relationship in (a) is obtained from weighted fitting.

F8. The storage medium of any one of embodiments F1 through F7, wherein each of the sequence reads for the sample is represented in binary format.

F9. The storage medium of embodiment F8, wherein the binary format for each read of the sequence comprises the chromosome to which the read maps and the chromosomal location to which the read maps.

F10. The storage medium of embodiment F9, wherein the binary format is in a 5-byte format comprising a 1-byte chromosomal ordinal number and a 4-byte chromosomal position.

F11. The storage medium of any one of embodiments F8-F10, wherein the binary format is 50 times smaller than a Sequence Alignment/Map (SAM) format and/or approximately 13% smaller than a GZip format.

F12. Any one of embodiments F1 through F11, wherein the normalizing in (c) comprises factorizing one or more features other than GC density and normalizing the sequence reads. storage media.

F13. The storage medium of embodiment F12, wherein the factorization of the one or more features is by a process that includes using a multivariate model.

F14. The storage medium of F13, wherein processing including using multivariate models is performed by a multivariate module.

F14.1. The storage medium of any one of embodiments F12-F14, wherein the sequence read counts are normalized according to the normalization in (c) and the factorization of the one or more features.

F15. The program causes the microprocessor to generate, after (c), a probability density estimate for each of the one or more portions comprising the counts of reads of the sequence normalized in (c). The storage medium of any one of embodiments F1 through F14.1 that directs to generate read densities for one or more portions of or segments thereof.

F16. The storage medium of embodiment F15, wherein the probability density estimates are kernel density estimates.

F17. The storage medium of embodiment F15 or F16, wherein the program directs the microprocessor to generate a read density profile for the genome or segments thereof.

F18. The storage medium of embodiment F17, wherein the read density profile comprises read densities for one or more portions of the genome or segments thereof.

F19. The storage medium of any one of embodiments F15-F18, wherein the program instructs the microprocessor to adjust each of the read densities for the one or more portions.

F20. The storage medium of any one of embodiments F15-F19, wherein one or more portions are filtered thereby providing filtered portions.

F21. The storage medium of any one of embodiments F15-F20, wherein the program instructs the microprocessor to weight one or more portions, thereby providing weighted portions.

F22. The storage medium of embodiment F21, wherein the one or more portions are weighted by eigenfunctions.

F23. The storage medium of any one of embodiments F1 through F22, wherein the program directs the microprocessor to obtain the sequence read prior to (a).

F24. The storage medium of embodiment F23, wherein the sequence reads are generated by massively parallel sequencing (MPS).

F25. The storage medium of embodiment F23 or F24, wherein the sequence reads obtained are sequence reads mapped to the entire reference genome or a segment of the genome.

F26. The storage medium of embodiment F25, wherein the segment of the genome comprises a chromosome or segment thereof.

F27. The storage medium of embodiment F25 or F26, wherein the read count of a sequence mapped against the reference genome is a normalized read count of the sequence.

F28. The storage medium of embodiment F27, wherein the read counts of sequences mapped against the reference genome are normalized by GC content, bin-wise normalization, GC LOESS, PERUN, GCRM, or combinations thereof.

F29. The storage medium of embodiment F25 or F26, wherein the read counts of sequences mapped against the reference genome are raw counts.

F30. The storage medium of any one of embodiments F15-F29, wherein each portion of the reference genome comprises approximately equal lengths of contiguous nucleotides.

F31. The storage medium of any one of embodiments F15 or F30, wherein each portion of the reference genome constitutes about 50 kb.

F32. The storage medium of any one of embodiments F15-F31, wherein each portion of the reference genome constitutes about 100 kb.

F33. The storage medium of any one of embodiments F15-F32, wherein each portion of the reference genome comprises a segment of contiguous nucleotides in common with adjacent portions of the reference genome.

F34. The storage medium of any one of embodiments F1-F33, wherein the test sample is obtained from a pregnant female.

F35. The storage medium of any one of embodiments F1-F34, wherein the test sample comprises blood from a pregnant female.

F36. The storage medium of any one of embodiments F1-F35, wherein the test sample comprises plasma from a pregnant female.

F37. The storage medium of any one of embodiments F1-F36, wherein the test sample comprises serum from a pregnant female.

F38. The storage medium of any one of embodiments F1 through F37, wherein the test sample comprises isolated nucleic acids.

F39. Any one of embodiments F8 through F38, wherein the program directs the microprocessor to compress the sequence reads mapped against the reference genome in (a) from sequence alignment format to binary format. storage media.

F40. The storage medium of embodiment F39, wherein the compression is performed by a compression module.

F41. The storage medium of any one of embodiments F1-F40, wherein the GC densities and GC density frequencies for the test sample sequence reads and for the references are provided by a biased density module.

F42. The storage medium of any one of embodiments F1-F41, wherein the comparison in (b) is generated by a relationship module.

F43. The storage medium of any one of embodiments F1-F42, wherein the normalization in (c) is performed by a bias correction module.

F44. The storage medium of any one of embodiments F15-F43, wherein the read density is provided by a distribution module.

F45. The storage medium of any one of embodiments F20-F44, wherein the filtered portion is provided by a filtering module.

F46. The storage medium of any one of embodiments F21-F45, wherein the weighted portion is provided by a portion weighting module.

F46.1. The storage medium of any one of embodiments F21-F45, wherein the adjusted read density is provided by a read density adjustment module.

F47. The storage medium of embodiment F46, comprising one or more of a compression module, a bias density module, a relation module, a bias correction module, a distribution module, a filtering module, a read density adjustment module, and a partial weighting module.

G1. A non-transitory computer-readable storage medium containing a self-stored executable program, the program stored in a microprocessor,
(a) filtering a portion of the reference genome according to the read density distribution, thereby providing a read density profile for the test sample comprising the read density of the filtered portion;
the read density includes reads of sequences of circulating cell-free nucleic acids from test samples from pregnant females;
a read density distribution is determined for the partial read densities for the plurality of samples;
(b) adjusting a read density profile for the test sample according to one or more principal components obtained from a set of known euploid samples by principal component analysis, thereby producing a test sample profile comprising the adjusted read density; to provide;
(c) comparing the test sample profile to the reference profile, thereby providing a comparison;

(d) determining the presence or absence of a chromosomal aneuploidy for the test sample according to the comparison;

G2. The storage medium of embodiment G1, wherein comparing includes determining a level of significance.

G3. The storage medium of embodiment G2, wherein determining the level of significance comprises determining a p-value.

G4. The storage medium of any one of embodiments G1-G3, wherein the reference profile comprises a read density profile obtained from a series of known euploid samples.

G5. The storage medium of any one of embodiments G1-G4, wherein the reference profile comprises read densities of the filtered portion.

G6. The storage medium of any one of embodiments G1-G5, wherein the reference profile comprises read densities adjusted according to one or more principal components.

G7. The storage of any one of embodiments G2 through G6, wherein the level of significance indicates a statistically significant difference between the test sample profile and the reference profile and the presence of a chromosomal aneuploidy is determined. media.

G8. The storage medium of any one of embodiments G1 through G7, wherein the plurality of samples comprises a series of known euploid samples.

G9. The storage medium of any one of embodiments G1 through G8, wherein the read density of the portion for the plurality of samples is the median read density.

G10. The storage medium of any one of embodiments G1 through G9, wherein the read density of the filtered portion for the test sample is the median read density.

G11. The storage medium of any one of embodiments G4-G10, wherein the read density profile for the reference profile comprises a median read density.

G12. The storage medium of any one of embodiments G4-G11, wherein read densities for the test sample profile, the plurality of samples, and the reference profile are determined according to a process that includes using kernel density estimation.

G13. The storage medium of any one of embodiments G10-G12, wherein the test sample profile is determined according to a median read density for the test samples.

G14. The storage medium of any one of embodiments G11-G13, wherein the reference profile is determined according to a median read density distribution for the references.

G15. The storage medium of any one of embodiments G1-G14, wherein the program directs the microprocessor to filter portions of the reference genome according to a measure of uncertainty about the read density distribution.

G15.1. The storage medium of embodiment G14.1, wherein the uncertainty measure is MAD.

G16. the program to the microprocessor,

(1) generating a fitted relationship between (i) guanine and cytosine (GC) density and (ii) GC density frequency for the sequence reads of the test sample, thereby generating a sample GC density relationship. the sequence reads are mapped against a reference genome;
(2) comparing the sample GC density relationship and the reference GC density relationship, thereby generating a comparison, comprising:
the reference GC density relationship is between (i) the GC density and (ii) the GC density frequency for the reference;

(3) the processing performed prior to (a), including normalizing the sequence read counts for the sample according to the comparison determined in (b), thereby reducing the sequence read bias for the sample; The storage medium of any one of embodiments G1 through G15.1, which directs weighting the read counts of the mapped sequence against the filtered portion for the test sample by .

G16.1. The storage medium of embodiment G16, wherein the normalizing in (3) includes providing a normalized count number.

G17. The storage medium of embodiment G16 or G16.1, wherein each of the GC densities is determined by a process that includes using kernel density estimation.

G17.1. The storage medium of any one of embodiments G16-G17, wherein each of the GC densities for the reference GC density relationship and the sample GC density relationship is an indication of local GC content.

G17.2. The storage medium of embodiment G17.1, wherein the local GC content is for polynucleotide segments of 5000 bp or less.

G18. The storage medium of any one of embodiments G16-G17.2, wherein each of the GC densities is determined by a process that includes using a sliding window analysis.

G19. The storage medium of embodiment G18, wherein the window is from about 5 contiguous nucleotides to about 5000 contiguous nucleotides and is slid from about 1 base to about 10 bases simultaneously in a sliding window analysis.

G20. The storage medium of embodiment G19, wherein the window is about 200 contiguous nucleotides and is slid about 1 base at a time in the sliding window analysis.

G21. (2) comprises generating (i) a ratio each comprising a sample GC density-related frequency and a reference GC density-related frequency for each of the GC densities and (ii) a fitted relationship with the GC density; The storage medium of any one of embodiments G16-G20.

G22. The storage medium of embodiment G21, wherein the fitted relationship in (1) is obtained from weighted fitting.

G23. The storage medium of any one of embodiments G16-G22, wherein each of the sequence reads for the sample is displayed in binary format.

G24. The storage medium of embodiment G23, wherein the binary format for each read of the sequence comprises the chromosome to which the read maps and the chromosomal location to which the read maps.

G25. The storage medium of embodiment G24, wherein the binary format is in a 5-byte format comprising a 1-byte chromosomal ordinal number and a 4-byte chromosomal position.

G26. The storage medium of any one of embodiments G23-G25, wherein the binary format is 50 times smaller than a sequence alignment/map (SAM) format and/or approximately 13% smaller than a GZip format.

G27. Any one of embodiments G16 through G26, wherein the normalizing in (c) comprises factorizing one or more features other than GC density and normalizing sequence read counts Storage media as described in .

G28. The storage medium of embodiment G27, wherein the factorization of the one or more features is by a process that includes using a multivariate model.

G29. The storage medium of embodiment G28, wherein the processing including using multivariate models is performed by a multivariate module.

G29.1. The storage medium of any one of embodiments G16-G29, wherein the program directs the microprocessor to weight the filtered portions for the test sample.

G29.2. The storage medium of embodiment G29.1, wherein the filtered portion for the test sample is weighted by a process that includes an eigenfunction.

G30. The storage medium of any one of embodiments G1 through G29.2, wherein the program instructs the microprocessor to obtain the sequence read prior to (a).

G31. The storage medium of embodiment G30, wherein the sequence reads are generated by massively parallel sequencing (MPS).

G32. The storage medium of any one of embodiments G1-G31, comprising obtaining sequence reads mapped to an entire reference genome or a segment of the genome.

G33. The storage medium of embodiment G32, wherein the segments of the genome comprise chromosomes or segments thereof.

G34. The storage medium of embodiment G32 or G33, wherein the read counts of the sequences mapped to the reference genome are normalized prior to (1).

G35. The storage medium of embodiment G34, wherein the read counts of the sequences mapped against the reference genome are normalized by GC content, bin-wise normalization, GC LOESS, PERUN, GCRM, or combinations thereof.

G36. The storage medium of embodiment G32 or G33, wherein the read counts of the sequences mapped against the reference genome are raw counts.

G37. The storage medium of any one of embodiments G1-G36, wherein each portion of the reference genome comprises approximately equal lengths of contiguous nucleotides.

G38. The storage medium of any one of embodiments G1-G37, wherein each portion of the reference genome constitutes about 50 kb.

G39. The storage medium of any one of embodiments G1-G38, wherein each portion of the reference genome constitutes approximately 100 kb.

G40. The storage medium of any one of embodiments G1-G39, wherein each portion of the reference genome comprises a segment of contiguous nucleotides in common with adjacent portions of the reference genome.

G41. The storage medium of any one of embodiments G1 through G40, wherein the test sample comprises blood from a pregnant female.

G42. The storage medium of any one of embodiments G1 through G41, wherein the test sample comprises plasma from a pregnant female.

G43. The storage medium of any one of embodiments G1 through G42, wherein the test sample comprises serum from a pregnant female.

G44 The storage medium of any one of embodiments G1 through G43, wherein nucleic acids are isolated from the test sample.

G45. Any one of embodiments G23 through G44, wherein the program directs the microprocessor to compress the sequence reads mapped to the reference genome in (1) from sequence alignment format to binary format. storage media.

G46. The storage medium of embodiment G45, wherein the compression is performed by a compression module.

G47. The storage medium of any one of embodiments G16-G46, wherein the GC densities and GC density frequencies for the test sample sequence reads and for the references are provided by a biased density module.

G48. The storage medium of any one of embodiments G16-G47, wherein the comparison in (2) is generated by a relationship module.

G49. The storage medium of any one of embodiments G17-G48, wherein the normalization in (3) is performed by a bias correction module.

G50. The storage medium of any one of embodiments G1-G49, wherein the read density is provided by a distribution module.

G51. The storage medium of any one of embodiments G1-G50, wherein the filtered portion is provided by a filtering module.

G51.1. The storage medium of any one of embodiments G29.1 through G51, wherein the filtered portions for the test sample are weighted by a portion weighting module.

G51.1. The storage medium of any one of embodiments G29.1-G51, wherein the adjusted read density is provided by a read density adjustment module.

G52. The method of embodiment G51.1, wherein the apparatus includes one or more of a compression module, a bias density module, a relation module, a bias correction module, a distribution module, a filtering module, a read density adjustment module, and a partial weighting module. storage media.

G53. The storage medium of any one of embodiments G1 through G52, wherein the test sample profile comprises a profile of chromosomes or segments thereof.

G54. The storage medium of any one of embodiments G1 through G53, wherein the reference profiles comprise profiles of chromosomes or segments thereof.

G55. The storage medium of any one of embodiments G1 through G54, wherein the determination in (d) is provided with a specificity equal to or greater than 90% and a sensitivity equal to or greater than 90%.

G56. The storage medium of any one of embodiments G1 through G55, wherein the aneuploidy is trisomy.

G57. The storage medium of embodiment G56, wherein the trisomy is trisomy 21, trisomy 18, or trisomy 13.

H1. A system comprising a memory and one or more microprocessors, wherein the one or more microprocessors are configured to perform processing according to instructions in the memory to reduce bias in sequence reads for a sample. and the processing is
(a) generating a relationship between (i) local genomic bias estimates and (ii) bias frequencies for sequence reads of a test sample, thereby generating a sample bias relationship, comprising:
the sequence reads are of circulating cell-free nucleic acid derived from the test sample;
the sequence reads are mapped against a reference genome;
(b) comparing the sample bias relationship and the reference bias relationship, thereby generating a comparison, comprising:
the reference bias relationship is between (i) a local genomic bias estimate and (ii) a bias frequency for the reference;

(c) normalizing the sequence read counts for the sample according to the comparisons determined in (b), thereby reducing sequence read bias for the sample.

H1.1. A system comprising a sequencing device and one or more computing devices,
The sequencing device is configured to generate a signal corresponding to the nucleotide bases of a nucleic acid loaded into the sequencing device, the nucleic acid being a circulating cell-free cell derived from the blood of a pregnant female giving birth to a fetus. the nucleic acid or the nucleic acid loaded into the sequencing device is a modified variant of the circulating cell-free nucleic acid;
The one or more computing units include memory and one or more processors, the memory including instructions executable by the one or more processors, the instructions executable by the one or more processors comprising:
Generate sequence reads from signals, map sequence reads,
(a) generating a relationship between (i) local genomic bias estimates and (ii) bias frequencies for the sequence reads of the test sample, thereby generating a sample bias relationship;
(b) comparing the sample-biased relation with the reference-biased relation, thereby generating a comparison (the reference-biased relation is between (i) the local genomic bias estimate and (ii) the biased frequency for the reference belongs to),

(c) a system configured to normalize the sequence read counts for the sample according to the comparison determined in (b), thereby reducing sequence read bias for the sample;

H1.2. The system of embodiment H1 or H1.1, wherein normalizing in (c) comprises providing a normalized count number.

H2. The system of embodiment H1 or H1.2, wherein each of the local genomic bias estimates is determined by a process that includes using kernel density estimation.

H2.1. The system of any one of embodiments H1-H2, wherein each of the local genomic bias estimates for the reference bias relationship and the sample bias relationship is an indication of local bias content.

H2.2. The system of embodiment H2.1, wherein the local bias content is for polynucleotide segments of 5000 bp or less.

H3. The system of any one of embodiments H1-H2.2, wherein each of the local genomic bias estimates is determined by a process that includes using a sliding window analysis.

H4. The system of embodiment H3, wherein the window is from about 5 consecutive nucleotides to about 5000 consecutive nucleotides and is slid from about 1 base to about 10 bases simultaneously in the sliding window analysis.

H5. The system of embodiment H3, wherein the window is about 200 contiguous nucleotides and is slid about 1 base at a time in the sliding window analysis.

H6. (b) is a ratio comprising (i) a sample-biased related frequency and a reference-biased related frequency, each for each of the local genomic bias estimates; and (ii) the local genomic bias estimate; The system as in any one of embodiments H1 through H5, comprising generating the fitted relationship of .

H7. The system of embodiment H6, wherein the fitted relationship in (a) is obtained from weighted fitting.

H8. The system of any one of embodiments H1 through H7, wherein each of the sequence reads for the sample is displayed in binary format.

H9. The system of embodiment H8, wherein the binary format for each read of the sequence comprises the chromosome to which the read maps and the chromosomal location to which the read maps.

H10. The system of embodiment H9, wherein the binary format is in a 5-byte format comprising a 1-byte chromosomal ordinal number and a 4-byte chromosomal position.

H11. The system of any one of embodiments H8 through H10, wherein the binary format is 50 times smaller than a sequence alignment/map (SAM) format and/or about 13% smaller than a GZip format.

H12. any one of embodiments H1 through H11, wherein the normalizing in (c) comprises factorizing one or more features other than bias and normalizing the sequence read counts System as described.

H13. The system of embodiment H12, wherein the factorization of the one or more features is by a process that includes using a multivariate model.

H14. The system of H13, wherein the processing including the use of multivariate models is performed by a multivariate module.

H14.1. The system of any one of embodiments H12-H14, wherein the sequence read counts are normalized according to the normalization in (c) and the factorization of the one or more features.

H15. (c) followed by generating a probability density estimate for each of the one or more portions comprising the normalized sequence read counts in (c); The system of any one of embodiments H1 through H14.1, comprising generating a read density for a portion or segment thereof.

H16. The system of embodiment H15, wherein the probability density estimation is kernel density estimation.

H17. The system of embodiment H15 or H16, comprising generating a read density profile for a genome or segment thereof.

H18. The system of embodiment H17, wherein the read density profile comprises read densities for one or more portions of the genome or segments thereof.

H19. The system of any one of embodiments H15-H18, comprising adjusting each of the read densities for the one or more portions.

H20. The system of any one of embodiments H15-H19, wherein one or more portions are filtered thereby providing filtered portions.

H21. The system of any one of embodiments H15-H20, wherein one or more portions are weighted thereby providing weighted portions.

H22. The system of embodiment H21, wherein the one or more portions are weighted by eigenfunctions.

H23. The system of any one of embodiments H1-H22, wherein the local genomic bias estimate comprises a local GC density and the bias frequency comprises a GC bias frequency.

H24. (a) filtering a portion of the reference genome according to the read density distribution, thereby providing a read density profile for the test sample comprising the read density of the filtered portion;
wherein the read density comprises reads of sequences of circulating cell-free nucleic acids from test samples derived from pregnant females;
a lead density distribution is determined for the partial lead densities for the plurality of samples;
(b) adjusting a read density profile for the test sample according to one or more principal components obtained from a set of known euploid samples by principal component analysis, thereby producing a test sample profile comprising the adjusted read density; providing a step;
(c) comparing the test sample profile to the reference profile, thereby providing a comparison;

(d) determining the presence or absence of a chromosomal aneuploidy for the test sample according to the comparison.

H25. The system of embodiment H24, wherein comparing includes determining a level of significance.

H26. The system of embodiment H25, wherein determining the level of significance comprises determining a p-value.

H27. The system of any one of embodiments H24-H26, wherein the reference profile comprises a read density profile obtained from a series of known euploid samples.

H28. The system of any one of embodiments H24-H27, wherein the reference profile comprises read densities of the filtered portion.

H29. The system of any one of embodiments H24-H28, wherein the reference profile comprises read densities adjusted according to one or more principal components.

H30. The system of any one of embodiments H25-H29, wherein the level of significance indicates a statistically significant difference between the test sample profile and the reference profile and the presence of a chromosomal aneuploidy is determined. .

H31. The system of any one of embodiments H24-H30, wherein the plurality of samples comprises a series of known euploid samples.

H32. The system of any one of embodiments H24 through H31, wherein the partial read density for the plurality of samples is the median read density.

H33. The system of any one of embodiments H24-H32, wherein the read density of the filtered portion for the test sample is the median read density.

H34. The system of any one of embodiments H27-H33, wherein the read density profile for the reference profile comprises a median read density.

H35. The system of any one of embodiments H27-H34, wherein read densities for the test sample profile, the plurality of samples, and the reference profile are determined according to a process that includes using kernel density estimation.

H36. The system of any one of embodiments H33-H35, wherein the test sample profile is determined according to a median read density for the test samples.

H37. The system of any one of embodiments H34-H36, wherein the reference profile is determined according to a median read density distribution for the reference.

H38. The system of any one of embodiments H24-H37, comprising filtering portions of the reference genome according to an uncertainty measure for the read density distribution.

H39. The system of embodiment H38, wherein the uncertainty measure is MAD.

H40. The system of any one of embodiments H1 through H39, wherein the memory of the system comprises reads of sequences of circulating cell-free nucleic acids from test samples that are mapped against a reference genome.

I1. A method for reducing sequence read bias for a sample, comprising:
(a) generating a relationship between (i) local genomic bias estimates and (ii) bias frequencies for sequence reads of a test sample using a microprocessor, thereby generating a sample bias relationship; is a step
the sequence reads are of circulating cell-free nucleic acid derived from the test sample;
the sequence reads are mapped against a reference genome;
(b) comparing the sample bias relationship and the reference bias relationship, thereby generating a comparison, comprising:
the reference bias relationship is between (i) a local genomic bias estimate and (ii) a bias frequency for the reference;

(c) normalizing the sequence read counts for the sample according to the comparison determined in (b), thereby reducing the sequence read bias for the sample.

I1.1. A method for reducing sequence read bias for a sample, comprising:
loading a circulating cell-free nucleic acid derived from the blood of a fetus-bearing pregnant female into a sequencing device, or loading a modified variant of the nucleic acid into a sequencing device, wherein the sequencing device generating a signal corresponding to the nucleotide base of
generating sequence reads from the nucleic acid signal, optionally after transfer of the signal to the system, by a system comprising one or more computing units, wherein the one or more computing units in the system: including memory and one or more processors;
one computing unit or combination of computing units in the system maps the sequence reads to the reference genome;
(a) using a microprocessor to generate a relationship between (i) local genomic bias estimates and (ii) bias frequencies for the sequence reads of the test sample, thereby generating a sample bias relationship; (The sequence reads are of circulating cell-free nucleic acids derived from the test sample,
Sequence reads are mapped against a reference genome),
(b) comparing the sample-biased relation to the reference-biased relation, thereby generating a comparison (the reference-biased relation is between (i) the local genomic bias estimate and (ii) the biased frequency for the reference; belongs to),

(c) normalizing the sequence read counts for the sample according to the comparison determined in (b), thereby reducing sequence read bias for the sample; Method.

I1.2. The method of embodiment I1 or I1.1, wherein normalizing in (c) comprises providing a normalized count number.

I2. The method of embodiment I1, I1.1 or I1.2, wherein the estimate of local genomic bias is determined by a process that includes the use of kernel density estimation.

I2.1. The method of any one of embodiments I1-I2, wherein each of the local genomic bias estimates for the reference bias relationship and the sample bias relationship is an indication of local bias content.

I2.2. The method of embodiment 12.1, wherein the local bias content is for polynucleotide segments of 5000 bp or less.

I3. The method of any one of embodiments 11-12.2, wherein each of the local genomic bias estimates is determined by a process that includes the use of a sliding window analysis.

I4. The method of embodiment I3, wherein the window is from about 5 contiguous nucleotides to about 5000 contiguous nucleotides and is slid from about 1 base to about 10 bases simultaneously in a sliding window analysis.

I5. The method of embodiment I3, wherein the window is about 200 contiguous nucleotides and is slid about 1 base at a time in the sliding window analysis.

I6. (b) is a ratio comprising (i) a sample-biased related frequency and a reference-biased related frequency, each for each of the local genomic bias estimates; and (ii) the local genomic bias estimate; The method as in any one of embodiments I1-I5, comprising generating the fitted relationship of .

I7. The method of embodiment I6, wherein the fitted relationship in (a) is obtained from weighted fitting.

I8. The method of any one of embodiments I1 through I7, wherein each of the sequence reads for the sample is displayed in binary format.

I9. The method of embodiment I8, wherein the binary format for each read of the sequence comprises the chromosome to which the read maps and the chromosomal location to which the read maps.

I10. The method of embodiment I9, wherein the binary format is in a 5-byte format comprising a 1-byte chromosomal ordinal number and a 4-byte chromosomal position.

I11. The method of any one of embodiments 18-110, wherein the binary format is 50 times smaller than a sequence alignment/map (SAM) format and/or about 13% smaller than a GZip format.

I12. any one of embodiments I1 through I11, wherein the normalizing in (c) comprises factorizing one or more features other than bias and normalizing the sequence read counts described method.

I13. The method of embodiment I12, wherein the factorization of the one or more features is by a process that includes using a multivariate model.

I14. The method of embodiment I13, wherein the processing including using the multivariate model is performed by a multivariate module.

I14.1. The method of any one of embodiments 112-114, wherein the sequence read counts are normalized according to the normalization in (c) and the factorization of the one or more features.

I15. (c) followed by generating a probability density estimate for each of the one or more portions comprising the normalized sequence read counts in (c); The method of any one of embodiments I1 through I14.1, comprising generating a read density for a portion or segment thereof.

I16. The method of embodiment 115, wherein the probability density estimation is kernel density estimation.

I17. The method of embodiment I15 or I16, comprising generating a read density profile for a genome or segment thereof.

I18. The method of embodiment I17, wherein the read density profile comprises read densities for one or more portions of the genome or segments thereof.

I19. The method of any one of embodiments I15-I18, comprising adjusting each of the lead densities for the one or more portions.

I20. The method of any one of embodiments I15-I19, wherein one or more portions are filtered, thereby providing filtered portions.

I21. The method of any one of embodiments I15-I20, wherein the one or more portions are weighted, thereby providing weighted portions.

I22. The method of embodiment I21, wherein the one or more portions are weighted by eigenfunctions.

I23. The method of any one of embodiments 11-122, wherein the local genomic bias estimate comprises a local GC density and the bias frequency comprises a GC bias frequency.

I23.1.
(a) filtering portions of the chromosomes in the reference genome according to the read density distribution, thereby providing a read density profile for the test sample comprising the read densities of the filtered portions;
the read density comprises reads of sequences of circulating cell-free nucleic acids from test samples from pregnant females;
a lead density distribution is determined for the partial lead densities for the plurality of samples;
(b) adjusting a chromosomal read density profile for the test sample according to one or more principal components obtained from a set of known euploid samples by principal component analysis, whereby the test sample comprising the adjusted read density; providing a chromosomal profile;
(c) comparing the test sample chromosomal profile to the reference profile, thereby providing a comparison;

(d) determining the presence or absence of a chromosomal aneuploidy for the test sample according to the comparison.

I24.
(a) filtering a portion of the reference genome according to the read density distribution, thereby providing a read density profile for the test sample comprising the read density of the filtered portion;
wherein the read density comprises reads of sequences of circulating cell-free nucleic acids from test samples derived from pregnant females;
a lead density distribution is determined for the partial lead densities for the plurality of samples;
(b) adjusting a read density profile for the test sample according to one or more principal components obtained from a set of known euploid samples by principal component analysis, thereby producing a test sample profile comprising the adjusted read density; providing a step;
(c) comparing the test sample profile to the reference profile, thereby providing a comparison;

(d) determining the presence or absence of a chromosomal aneuploidy for the test sample according to the comparison.

I24.1. The method of embodiment I23.1 or I24, wherein the read density profile is adjusted in (b) by 1-10 principal components.

I24.2. The method of embodiment I23.1, I24, or I24.1, wherein the read density profile is adjusted in (b) by five principal components.

I24.3. One or more principal components adjusted for one or more features in the read density profile, the features correlated to fetal sex, sequence bias, fetal fraction, DNase I sensitivity bias , entropy, repetitive sequence bias, chromatin structure bias, polymerase error rate bias, batch sequence bias, inverted repeat bias, PCR amplification bias, and hidden copy number variation, embodiment I23. .The method of any one of 1 to I24.2.

I24.4. The method of embodiment 124.3, wherein the sequence bias comprises a guanine and cytosine (GC) bias.

I25. The method of any one of embodiments I23.1 through I24.4, wherein comparing comprises determining a level of significance.

I26. The method of embodiment 125, wherein determining the level of significance comprises determining a p-value.

I27. The method of any one of embodiments 123.1-126, wherein the reference profile comprises a read density profile obtained from a series of known euploid samples.

I28. The method of any one of embodiments 123.1-127, wherein the reference profile comprises read densities of the filtered portion.

I29. The method of any one of embodiments 123.1-128, wherein the reference profile comprises read densities adjusted according to one or more principal components.

I30. The method of any one of embodiments I25-I29, wherein the level of significance indicates a statistically significant difference between the test sample profile and the reference profile and the presence of a chromosomal aneuploidy is determined. .

I31. The method of any one of embodiments 123.1-130, wherein the plurality of samples comprises a series of known euploid samples.

I32. The method of any one of embodiments 123.1 through 131, wherein the partial lead density for the plurality of samples is the median lead density.

I33. The method of any one of embodiments 123.1 through 132, wherein the read density of the filtered portion for the test sample is the median read density.

I34. The method of any one of embodiments 127-133, wherein the read density profile for the reference profile comprises a median lead density.

I35. The method of any one of embodiments 127-134, wherein read densities for the test sample profile, the plurality of samples, and the reference profile are determined according to a process that includes using kernel density estimation.

I36. The method of any one of embodiments 133-135, wherein the test sample profile is determined according to the median read density for the test samples.

I37. The method of any one of embodiments 134-136, wherein the reference profile is determined according to a median read density distribution for the reference.

I38. The method of any one of embodiments I23.1 through I37, comprising filtering portions of the reference genome according to an uncertainty measure for the read density distribution.

I39. The method of embodiment I38, wherein the uncertainty measure is MAD.

I40. The method of any one of embodiments I23.1 through I39, wherein the test sample profile is a representation of chromosomal dose for the test sample.

I41. The method of embodiment 140, comprising comparing the chromosome dose for the test sample profile to the chromosome dose for the reference profile, thereby generating a chromosome dose comparison.

I42. The method of embodiment I41, wherein determining the presence or absence of a chromosomal aneuploidy for the test sample follows a comparison of chromosome dose.

I43. A determination of the presence or absence of a chromosomal aneuploidy for a test sample results in 1 copy of the chromosome, 2 copies of the chromosome, 3 copies of the chromosome, 4 copies of the chromosome, 5 copies of the chromosome, 1 copy of the chromosome The method of embodiment I42, comprising identifying the presence or absence of deletions of one or more segments or insertions of one or more segments of the chromosome.

J1. A method for determining the presence or absence of aneuploidy comprising:
(a) obtaining a count of partial nucleotide sequence reads mapped to a genomic portion of a reference genome, wherein the partial nucleotide sequence reads are in test samples from pregnant females; Circulating cell-free nucleic acid reads from which at least a portion of the partial nucleotide sequence reads comprise
i) a plurality of nucleobase gaps between identified nucleobases, or ii) one or more nucleobase classes, each comprising a subset of the nucleobases present in the sample nucleic acid, or (i) and (ii) ), and
(b) filtering a portion of the reference genome according to the read density distribution, thereby providing a read density profile for the test sample comprising read densities of the filtered portion;
The read density includes partial nucleotide sequence reads from the test sample,
a lead density distribution is determined for the partial lead densities for the plurality of samples;
(c) adjusting a read density profile for the test sample according to one or more principal components obtained from a set of known euploid samples by principal component analysis, thereby producing a test sample profile comprising the adjusted read densities; providing a step;
(d) comparing the test sample profile to the reference profile, thereby providing a comparison;

(e) determining the presence or absence of aneuploidy for the test sample according to the comparison.

J2. A method for determining a fetal fraction based on copy number variation, comprising:
(a) obtaining a count of nucleic acid sequence reads mapped to a genomic portion of a reference genome, wherein the sequence reads are circulating cell-free nucleic acids derived from a test sample derived from a pregnant female; a step, which is the lead of
(b) normalizing the mapped counts to genomic portions of the reference genome, thereby providing normalized counts for the genomic portions, comprising:

(i) filtering a portion of the reference genome according to the read density distribution, thereby providing a read density profile for the test sample comprising the read density of the filtered portion;
The read density includes nucleotide sequence reads derived from the test sample,
a read density distribution is determined for the partial read densities for the plurality of samples;

(ii) adjusting a read density profile for the test sample according to one or more principal components obtained from a set of known euploid samples by principal component analysis, thereby producing a test sample profile comprising the adjusted read density; providing;

(c) identifying a first level of normalized counts that is significantly different from a second level of normalized counts, wherein the first level is a first set of genomic portions; and the second level is for a second set of genome portions;
(d) assigning copy number variations to a first level, thereby providing a classification;

(e) determining the fetal fraction of circulating cell-free nucleic acids according to the classification, whereby the fetal fraction is generated from the nucleic acid sequence reads.

J3. 1. A method for determining the fraction of fetal nucleic acid in circulating cell-free nucleic acid from the blood of a pregnant female, comprising:
(a) obtaining a count of nucleic acid sequence reads mapped to a genomic portion of a reference genome, wherein the sequence reads are derived from a test sample derived from a pregnant female giving birth to a male fetus; is a circulating cell-free nucleic acid lead that
(b) generating an experimental X-chromosome representation, the experimental X-chromosome representation comprising: (i) a read count of the sequence mapped to the genomic portion of the reference genome in the X chromosome; , (ii) the ratio of the read count of the sequence mapped to the genome portion of the reference genome or segment thereof in the genome;

(c) determining, from the experimental X chromosome representation, the fraction of fetal nucleic acids in the blood of the pregnant female according to the experimental X chromosome representation and the predicted X chromosome representation, wherein: The representation is the ratio of (i) the number of genomic parts of the reference genome in the X chromosome to (ii) the number of genomic parts of the reference genome or segments thereof in the genome, where the count in (b) is

(1) filtering a portion of the reference genome according to the read density distribution, thereby providing a read density profile for the test sample comprising the read density of the filtered portion;
The read density includes nucleotide sequence reads derived from the test sample,
a read density distribution is determined for the partial read densities for the plurality of samples;

(2) adjusting a read density profile for the test sample according to one or more principal components obtained from a set of known euploid samples by principal component analysis, thereby producing a test sample profile comprising the adjusted read density; and normalized by a process comprising providing.

J4. 1. A method for determining fetal ploidy according to a nucleic acid sequence read, comprising:
(a) determining a fraction of fetal nucleic acids in a test sample, wherein the test sample comprises circulating cell-free nucleic acids from pregnant females;
(b) obtaining a count of sequence reads mapped to a portion of the reference genome, wherein the sequence reads are derived from nucleic acids in the sample;
(c) calculating a level of genome division for each of the portions of the reference genome, thereby providing the calculated level of genome division;
(d) determining fetal ploidy according to the relationship between (i) the level of genomic division calculated for a subset of the portion of the reference genome and (ii) the fraction of fetal nucleic acid determined in (a); , the number of counts in (b) is

(1) filtering a portion of the reference genome according to the read density distribution, thereby providing a read density profile for the test sample comprising the read density of the filtered portion;
The read density includes nucleotide sequence reads derived from the test sample,
a read density distribution is determined for the partial read densities for the plurality of samples;

(2) adjusting a read density profile for the test sample according to one or more principal components obtained from a set of known euploid samples by principal component analysis, thereby producing a test sample profile comprising the adjusted read density; and normalized by a process comprising providing.

J5. A method for determining the presence or absence of fetal aneuploidy comprising:
(a) obtaining a count of nucleotide sequence reads mapped to a reference genome portion, wherein the nucleotide sequence reads are circulating cell-free nucleic acids derived from a test sample derived from a pregnant female; a step that is a lead;
(b) a process comprising subtracting the expected number of counts from the number of counts for the first genome portion, thereby producing a subtracted value, and dividing the subtracted value by an estimate of the variability of the counts; or using a microprocessor to normalize the derivative of the counts for the first genome portion, thereby obtaining a normalized sample count. and
the expected counts, or a derivative of the expected counts, is obtained for the sample, the reference, or the group comprising the sample and the reference exposed to one or more common experimental conditions;
(c) determining the presence or absence of fetal aneuploidy based on the normalized sample counts, wherein the normalization of the counts in (b) comprises:

(1) filtering a portion of the reference genome according to the read density distribution, thereby providing a read density profile for the test sample comprising the read density of the filtered portion;
The read density includes nucleotide sequence reads derived from the test sample,
a read density distribution is determined for the partial read densities for the plurality of samples;

(2) adjusting a read density profile for the test sample according to one or more principal components obtained from a set of known euploid samples by principal component analysis, thereby producing a test sample profile comprising the adjusted read density; and providing.

J6. A method for determining the sex chromosome karyotype in a fetus comprising:
(a) obtaining a count of nucleotide sequence reads mapped to a portion of a reference genome, wherein the sequence reads are of circulating cell-free nucleic acid derived from a test sample derived from a pregnant female; a step that is a lead;
(b) a plurality of samples from the matched relationship for each sample of (i) sequence read counts mapped to each of the portions of the reference genome and (ii) mapping features for each of the portions; determining an experimental bias for each of the portions of the reference genome for
(c) calculating a level of genome division for each of the portions of the reference genome from the fitted relationship between the experimental bias and counts of sequence reads mapped to each of the portions of the reference genome; providing the level of genome division calculated by
(d) determining the sex chromosome karyotype for the fetus according to the calculated level of genomic division, wherein determining the experimental bias in (b) comprises:

(1) filtering a portion of the reference genome according to the read density distribution, thereby providing a read density profile for the test sample comprising the read density of the filtered portion;
The read density includes nucleotide sequence reads derived from the test sample,
a read density distribution is determined for the partial read densities for the plurality of samples;

(2) adjusting a read density profile for the test sample according to one or more principal components obtained from a set of known euploid samples by principal component analysis, thereby producing a test sample profile comprising the adjusted read density; and providing.

J7. A method for determining the presence or absence of aneuploidy comprising:
(a) Obtaining a count of sequence reads mapped to chromosomes 13, 18, and 21, or segments thereof, in the reference genome, wherein the sequence reads are from pregnant females tested is a read of circulating cell-free nucleic acid derived from the sample;
(b) determining three ratios or ratio values, each of the three ratios being: (i) the number of counts mapped to each of chromosomes 13, 18, and 21, or segments thereof; (ii) is the ratio of the number of counts mapped to each of the other chromosomes 13, 18, and 21, or segments thereof;
(c) comparing three ratios or ratio values, thereby producing a comparison;

(d) the comparisons made in (c) and the determinations in (d) are not based on segments of the genome other than those in chromosomes 13, 18 and 21; determining the presence or absence of a chromosomal aneuploidy based on, whereby the determination of the presence or absence of a chromosomal aneuploidy is generated from the sequence reads, comprising: chromosomes 13, 18 and 21; or counts of sequence reads mapped to that segment,

(1) filtering a portion of the reference genome according to the read density distribution, thereby providing a read density profile for the test sample comprising the read density of the filtered portion;
The read density includes nucleotide sequence reads derived from the test sample,
a read density distribution is determined for the partial read densities for the plurality of samples;

(2) adjusting a read density profile for the test sample according to one or more principal components obtained from a set of known euploid samples by principal component analysis, thereby producing a test sample profile comprising the adjusted read density; and normalized by a process comprising providing.

Each patent, patent application, publication, and document referenced herein is hereby incorporated by reference in its entirety. Citation of any of the above patents, patent applications, publications, and documents is not an admission that any of the above materials are pertinent prior art, nor is there any admission as to the contents or dates of these publications or documents. It's not even something to be.

Modifications may be made to the above without departing from the basic aspects of the technology. The present technology has been described in considerable detail with reference to one or more specific embodiments, and persons skilled in the art may make modifications to the embodiments specifically disclosed in this application. As will be appreciated, these modifications and improvements remain within the scope and spirit of the present technology.

The technology illustratively described herein may suitably be practiced without any of the element(s) not specifically disclosed herein. Thus, for example, in each instance herein, any of the terms "comprising," "consisting essentially of," and "consisting of" the other The two terms are interchangeable. The terms and phrases employed are used as terms of description rather than of limitation, and the use of such terms and phrases does not exclude any property shown or described, or a segment equivalent thereof, Various modifications are possible within the scope of the claimed technology. The term "a" or "an" unless it is clear from the context that one of the elements, or more than one of the elements, is being described, it modifies It may refer to one or more elements (eg, "a reagent" may mean one or more reagents). The term "about," as used herein, refers to values within 10% of the underlying parameter (e.g., plus or minus 10%), and the term "about" at the beginning of a range of values. When used, the term modifies each of the values (eg, "about 1, 2, and 3" refers to about 1, about 2, and about 3). For example, a weight of "about 100 grams" can include weights between 90 grams and 110 grams. Further, when a list of values is described herein (e.g., about 50%, 60%, 70%, 80%, 85%, or 86%), the list includes all intermediate values and fractions thereof. (eg, 54%, 85.4%) are included. Thus, while the technology has been specifically disclosed by way of representative embodiments and optional features, it should be understood that modifications and variations of the concepts disclosed herein can be implemented by those skilled in the art. and such modifications and variations are deemed to be within the scope of this technology.

Certain embodiments of the technology are set forth in the claim(s) that follow.
For example, the present invention provides the following items.
(Item 1)
A method for determining the presence or absence of aneuploidy for a sample, comprising:
(a) filtering a portion of the reference genome according to the read density distribution, thereby providing a read density profile for the test sample comprising the read density of the filtered portion;
the read density comprises reads of sequences of circulating cell-free nucleic acids from test samples derived from pregnant females;
wherein the lead density distribution is determined for partial lead densities for a plurality of samples;
(b) adjusting said read density profile for said test sample according to one or more principal components obtained from a set of known euploid samples by principal component analysis, thereby comprising an adjusted read density; providing a profile;
(c) comparing said test sample profile to a reference profile, thereby providing a comparison;
(d) determining the presence or absence of a chromosomal aneuploidy for said test sample according to said comparison.
(Item 2)
The method of item 1, wherein the adjusting in (b) is performed using a microprocessor.
(Item 3)
A method for determining the presence or absence of aneuploidy for a sample, comprising:
loading a circulating cell-free nucleic acid derived from the blood of a fetus-bearing pregnant female into a sequencing device, or loading a modified variant of said nucleic acid into said sequencing device, said sequencing device produces a signal corresponding to the nucleotide bases of said nucleic acid;
generating sequence reads from said signal of said nucleic acid, optionally after transferring said signal to a system, by a system comprising one or more computing units, said one or more in said system The computing unit of includes memory and one or more processors,
A computing unit or combination of computing units in the system maps the reads of the sequence to a reference genome;
(a) filtering a portion of the reference genome according to the read density distribution, thereby providing a read density profile for the test sample comprising read densities of the filtered portion, said read densities from pregnant females comprising sequence reads of circulating cell-free nucleic acids derived from the test sample;
the lead density distribution is determined for partial lead densities for a plurality of samples;
(b) using a microprocessor to adjust said read density profile for said test sample according to one or more principal components obtained from a series of known euploid samples by principal component analysis; provide a test sample profile including read density,
(c) comparing said test sample profile to a reference profile, thereby providing a comparison;
(d) configured to determine the presence or absence of a chromosomal aneuploidy for said test sample according to said comparison.
(Item 4)
4. The method of items 1, 2, or 3, wherein the read density profile is adjusted in (b) by 1-10 principal components.
(Item 5)
4. The method of items 1, 2, or 3, wherein the read density profile is adjusted in (b) by five principal components.
(Item 6)
wherein said one or more principal components adjust for one or more features in a read density profile, wherein said features are fetal sex, sequence bias, fetal fraction, bias correlated to DNase I sensitivity, entropy , repeat bias, chromatin structure bias, polymerase error rate bias, batch sequence bias, inverted repeat bias, PCR amplification bias, and hidden copy number variation. A method according to any one of paragraphs.
(Item 7)
7. The method of item 6, wherein the sequence bias comprises a guanine and cytosine (GC) bias.
(Item 8)
8. The method of any one of items 1-7, wherein said comparing comprises determining a level of significance.
(Item 9)
9. The method of any one of items 1-8, wherein determining the level of significance comprises determining a p-value.
(Item 10)
10. The method of any one of items 1-9, wherein the reference profile comprises a read density profile obtained from a series of known euploid samples.
(Item 11)
11. The method of any one of items 1-10, wherein the reference profile comprises read densities of a filtered portion.
(Item 12)
12. The method of any one of items 1-11, wherein the reference profile comprises read densities adjusted according to the one or more principal components.
(Item 13)
13. Any one of items 8 to 12, wherein said level of significance indicates a statistically significant difference between said test sample profile and said reference profile and said presence of a chromosomal aneuploidy is determined. described method.
(Item 14)
14. The method of any one of items 1-13, wherein the plurality of samples comprises a series of known euploid samples.
(Item 15)
15. The method of any one of items 1-14, wherein the lead density of a portion for the plurality of samples is a median lead density.
(Item 16)
16. The method of any one of items 1-15, wherein the read density of the filtered portion for the test sample is a median read density.
(Item 17)
17. The method of any one of items 10-16, wherein the read density profile for the reference profile comprises a median lead density.
(Item 18)
18. The method of any one of items 10-17, wherein the read densities for the test sample profile, the plurality of samples and the reference profile are determined according to a process comprising using kernel density estimation.
(Item 19)
19. The method of any one of items 16-18, wherein the test sample profile is determined according to the median read density for the test samples.
(Item 20)
20. The method of any one of items 17-19, wherein the reference profile is determined according to the median read density distribution for the references.
(Item 21)
21. The method of any one of items 1-20, comprising filtering portions of the reference genome according to an uncertainty measure for the read density distribution.
(Item 22)
22. The method of item 21, wherein the uncertainty measure is MAD.
(Item 23)
23. The method of any one of items 1-22, wherein the test sample profile is an indication of chromosomal dose for the test sample.
(Item 24)
24. The method of item 23, comprising comparing the chromosome dose for the test sample profile to the chromosome dose for the reference profile, thereby generating a chromosome dose comparison.
(Item 25)
25. The method of item 24, wherein determining the presence or absence of a chromosomal aneuploidy for the test sample is according to a comparison of the chromosome dose.
(Item 26)
Determining the presence or absence of a chromosomal aneuploidy for the test sample results in: 1 copy of the chromosome, 2 copies of the chromosome, 3 copies of the chromosome, 4 copies of the chromosome, 5 copies of the chromosome, 26. A method according to any one of items 1 to 25, comprising identifying the presence or absence of deletions of one or more segments of or insertions of one or more segments of a chromosome.
(Item 27)
a read count of said sequence mapped to a filtered portion for said test sample,
(I) generating a relationship between (i) local genomic bias estimates and (ii) bias frequencies for sequence reads of a test sample, thereby generating a sample bias relationship,
wherein said sequence reads are of circulating cell-free nucleic acid derived from said test sample;
the reads of said sequence are mapped against a reference genome;
(II) comparing said sample bias relationship and a reference bias relationship, thereby generating a comparison, comprising:
said reference bias relationship is between (i) a local genomic bias estimate and (ii) said bias frequency for a reference;
(III) normalizing the sequence read counts for the sample according to the comparison determined in (II), thereby reducing the sequence read bias for the sample (a 27. A method according to any one of items 1 to 26, wherein normalization is performed by processing performed prior to ).
(Item 28)
28. The method of item 27, wherein the normalizing in (III) comprises providing normalized counts.
(Item 29)
29. The method of item 27 or 28, wherein each of the local genomic bias estimates is determined by a process comprising using kernel density estimation.
(Item 30)
30. The method of any one of items 27-29, wherein each of the local genomic bias estimates for the reference bias relationship and the sample bias relationship is an indication of local bias content. .
(Item 31)
31. The method of item 30, wherein the local bias content is for polynucleotide segments of 5000 bp or less.
(Item 32)
32. The method of any one of items 27-31, wherein each of the local genomic bias estimates is determined by a process comprising the use of a sliding window analysis.
(Item 33)
33. The method of item 32, wherein the window is from about 5 consecutive nucleotides to about 5000 consecutive nucleotides and is slid from about 1 base to about 10 bases simultaneously in the sliding window analysis.
(Item 34)
33. The method of item 32, wherein the window is about 200 consecutive nucleotides and is slid about 1 base at a time in the sliding window analysis.
(Item 35)
(II) is a ratio comprising (i) a sample-biased related frequency and a reference-biased related frequency, each for each of said local genomic bias estimates, and (ii) a local genomic bias estimate; 35. A method according to any one of items 27 to 34, comprising generating a fitted relationship with
(Item 36)
36. The method of item 35, wherein the fitted relationship in (I) is obtained from weighted fitting.
(Item 37)
37. The method of any one of items 27-36, wherein each of said sequence reads for said sample is displayed in a binary format.
(Item 38)
38. The method of item 37, wherein said binary format for each read of said sequence comprises a chromosome to which said read maps and a chromosomal location to which said read maps.
(Item 39)
39. The method of item 38, wherein the binary format is in a 5-byte format comprising a 1-byte chromosomal ordinal number and a 4-byte chromosomal position.
(Item 40)
40. The method of any one of items 37-39, wherein said binary format is 50 times smaller than a Sequence Alignment/Map (SAM) format and/or about 13% smaller than a GZip format.
(Item 41)
41. Any one of items 27-40, wherein said normalizing in (III) comprises factorizing one or more features other than bias and normalizing read counts of said sequence The method described in .
(Item 42)
42. The method of item 41, wherein said factorization of one or more features is by a process that includes using a multivariate model.
(Item 43)
43. The method of 42, wherein said processing including use of said multivariate model is performed by a multivariate module.
(Item 44)
44. The method of any one of items 41-43, wherein the sequence read counts are normalized according to the normalization in (III) and the factorization of the one or more features.
(Item 45)
(III), followed by generating a probability density estimate for each of the one or more portions comprising one or more of the sequence read counts normalized in (III), of the genome. 45. The method of any one of items 27-44, comprising generating a read density for one or more portions or segments thereof.
(Item 46)
46. The method of item 45, wherein the probability density estimate is a kernel density estimate.
(Item 47)
47. A method according to item 45 or 46, comprising generating a read density profile for said genome or said segment thereof.
(Item 48)
48. The method of item 47, wherein said read density profile comprises said read density for said one or more portions of said genome or said segments thereof.
(Item 49)
49. The method of any one of items 45-48, comprising adjusting each of said lead densities for said one or more portions.
(Item 50)
50. A method according to any one of items 45 to 49, wherein said one or more portions are filtered thereby providing filtered portions.
(Item 51)
51. The method of any one of items 45-50, wherein the one or more portions are weighted thereby providing weighted portions.
(Item 52)
52. The method of item 51, wherein the one or more portions are weighted by eigenfunctions.
(Item 53)
53. The method of any one of items 27-52, wherein the local genomic bias estimate is the local GC density and the bias frequency is the GC bias frequency.
(Item 54)
a read count of said sequence mapped to a filtered portion for said test sample,
(1) generating a fitted relationship between (i) guanine and cytosine (GC) density and (ii) GC density frequency for reads of said sequence of said test sample, thereby generating a sample GC density relationship; wherein the sequence reads are mapped against the reference genome;
(2) comparing the sample GC density relationship with a reference GC density relationship, thereby generating a comparison;
said reference GC density relationship is between (i) a GC density and (ii) said GC density frequency for a reference;
(3) normalizing the sequence read counts for the sample according to the comparison determined in (2), thereby reducing the sequence read bias for the sample; 27. A method according to any one of items 1 to 26, wherein normalization is performed by processing performed prior to .
(Item 55)
55. The method of item 54, wherein the normalizing in (3) includes providing normalized counts.
(Item 56)
56. Method according to item 54 or 55, wherein each of said GC densities is determined by a process comprising using kernel density estimation.
(Item 57)
57. The method of any one of items 54-56, wherein each of said GC densities is determined by a process that includes using a sliding window analysis.
(Item 58)
58. The method of item 57, wherein the window is from about 5 consecutive nucleotides to about 5000 consecutive nucleotides and is slid from about 1 base to about 10 bases simultaneously in the sliding window analysis.
(Item 59)
59. The method of item 58, wherein said window is about 200 contiguous nucleotides and is slid about 1 base at a time in said sliding window analysis.
(Item 60)
(2) comprises generating (i) a ratio each comprising a sample GC density-related frequency and a reference GC density-related frequency for each of said GC densities, and (ii) a fitted relationship with GC density; , items 54 to 59.
(Item 61)
61. The method of item 60, wherein the fitted relationship in (1) is obtained from weighted fitting.
(Item 62)
62. The method of any one of items 54-61, wherein each of said sequence reads for said sample is displayed in a binary format.
(Item 63)
63. The method of item 62, wherein the binary format for each read of the sequence comprises a chromosome to which the read is mapped and a chromosomal location to which the read is mapped.
(Item 64)
64. The method of item 63, wherein the binary format is in a 5-byte format comprising a 1-byte chromosomal ordinal number and a 4-byte chromosomal position.
(Item 65)
65. The method of any one of items 62-64, wherein said binary format is 50 times smaller than a Sequence Alignment/Map (SAM) format and/or approximately 13% smaller than a GZip format.
(Item 66)
66. The method of any one of items 54-65, wherein said normalizing in (c) comprises factorizing one or more features other than GC density and normalizing reads of said sequence. the method of.
(Item 67)
67. The method of item 66, wherein said factorization of one or more features is by a process that includes using a multivariate model.
(Item 68)
68. The method of item 67, wherein said processing comprising use of said multivariate model is performed by a multivariate module.
(Item 69)
69. The method of any one of items 54-68, wherein the filtered portion for the test sample is weighted.
(Item 70)
70. The method of item 69, wherein the filtered portion for the test sample is weighted by a process comprising eigenfunctions.
(Item 71)
71. The method of any one of items 1-70, comprising obtaining reads of said sequence prior to (a).
(Item 72)
72. The method of item 71, wherein the sequence reads are generated by Massively Parallel Sequencing (MPS).
(Item 73)
73. A method according to any one of items 1 to 72, comprising obtaining sequence reads mapped to the whole of the reference genome or to a segment of the genome.
(Item 74)
74. The method of item 73, wherein said segment of said genome comprises a chromosome or segment thereof.
(Item 75)
75. The method of item 73 or 74, wherein the count of reads of the sequence mapped against the reference genome is normalized prior to (1).
(Item 76)
76. The method of item 75, wherein the count of reads of the sequence mapped against the reference genome is normalized by GC content, bin-wise normalization, GC LOESS, PERUN, GCRM, or combinations thereof. .
(Item 77)
75. The method of item 73 or 74, wherein the count of reads of the sequence mapped against the reference genome is a raw count.
(Item 78)
78. The method of any one of items 1-77, wherein each portion of the reference genome comprises approximately equal lengths of contiguous nucleotides.
(Item 79)
79. The method of any one of items 1-78, wherein each portion of the reference genome constitutes approximately 50 kb.
(Item 80)
79. The method of any one of items 1-78, wherein each portion of the reference genome constitutes approximately 100 kb.
(Item 81)
81. The method of any one of items 1-80, wherein each portion of the reference genome comprises a segment of contiguous nucleotides in common with adjacent portions of the reference genome.
(Item 82)
82. The method of any one of items 1-81, wherein the test sample comprises blood from a pregnant female.
(Item 83)
82. The method of any one of items 1-81, wherein the test sample comprises plasma from a pregnant female.
(Item 84)
82. The method of any one of items 1-81, wherein the test sample comprises serum from a pregnant female.
(Item 85)
85. The method of any one of items 1-84, wherein nucleic acids are isolated from the test sample.
(Item 86)
86. The method of any one of items 62-85, comprising compressing the reads of the sequence mapped against the reference genome in (1) from a sequence alignment format to a binary format.
(Item 87)
87. Method according to item 86, wherein said compression is performed by a compression module.
(Item 88)
88. The method of any one of items 54-87, wherein the GC density and the GC density frequency for the sequence reads of the test sample and for the reference are provided by a biased density module.
(Item 89)
88. The method of any one of items 54-87, wherein the comparison in (2) is generated by a relationship module.
(Item 90)
90. The method of any one of items 54-89, wherein the normalization in (3) is performed by a bias correction module.
(Item 91)
91. The method of any one of items 1-90, wherein the lead density is provided by a distribution module.
(Item 92)
92. Method according to any one of items 1 to 91, wherein the filtered portion is provided by a filtering module.
(Item 93)
93. The method of any one of items 69-92, wherein the filtered portions of the test sample are weighted by a portion weighting module.
(Item 94)
94. The method of any one of items 69-93, wherein the lead density is adjusted by a lead density adjustment module.
(Item 95)
Items 1-94 performed using an apparatus including one or more of a compression module, a bias density module, a relation module, a bias correction module, a distribution module, a filtering module, a read density adjustment module, and a partial weighting module. The method according to any one of .
(Item 96)
96. The method of any one of items 1-95, wherein said test sample profile comprises a profile of a chromosome or a segment thereof.
(Item 97)
97. Method according to any one of items 1 to 96, wherein said reference profile comprises a profile of a chromosome or a segment thereof.
(Item 98)
98. The method of any one of items 1-97, wherein said determining in (d) is provided with a specificity equal to or greater than 90% and a sensitivity equal to or greater than 90%.
(Item 99)
99. The method of any one of items 1-98, wherein said aneuploidy is a trisomy.
(Item 100)
100. The method of item 99, wherein the trisomy is trisomy 21, trisomy 18, or trisomy 13.
(Item 101)
A method for determining the presence or absence of aneuploidy for a sample, comprising:
(a) normalizing the sequence read counts mapped to a reference genome, wherein the sequence reads are circulating cell-free nucleic acid derived from a test sample derived from a pregnant female; and the normalization is:
(1) generating a fitted relationship between (i) guanine and cytosine (GC) density and (ii) GC density frequency for reads of said sequence of said test sample, thereby generating a sample GC density relationship; wherein the sequence reads are mapped against the reference genome;
(2) comparing the sample GC density relationship with a reference GC density relationship, thereby generating a comparison;
said reference GC density relationship is between (i) a GC density and (ii) said GC density frequency for a reference;
(3) normalizing the sequence read counts for the sample according to the comparison determined in (2), thereby reducing the sequence read bias for the sample;
(b) filtering portions of the reference genome according to a read density distribution, thereby providing a read density profile for the test sample comprising the read densities of the filtered portions;
the read density comprises reads of sequences derived from the test sample;
wherein the lead density distribution is determined for partial lead densities for a plurality of samples;
(c) adjusting said read density profile for said test sample according to one or more principal components obtained from a set of known euploid samples by principal component analysis, whereby a test sample comprising an adjusted read density; providing a profile;
(d) comparing said test sample profile to a reference profile, thereby providing a comparison;
(e) determining the presence or absence of a chromosomal aneuploidy for said test sample according to said comparison.
(Item 102)
A method for reducing sequence read bias for a sample, comprising:
(a) generating a relationship between (i) local genomic bias estimates and (ii) bias frequencies for sequence reads of a test sample, thereby generating a sample bias relationship, comprising:
said sequence reads are of circulating cell-free nucleic acid derived from said test sample;
the reads of said sequence are mapped against a reference genome;
(b) comparing the sample bias relationship and a reference bias relationship, thereby generating a comparison, comprising:
said reference bias relationship is between (i) a local genomic bias estimate and (ii) said bias frequency for a reference;
(c) normalizing the sequence read counts for the sample according to the comparison determined in (b), thereby reducing the sequence read bias for the sample. .
(Item 103)
103. The method of item 102, wherein the sample bias relationship is generated using a microprocessor.
(Item 104)
A method for reducing sequence read bias for a sample, comprising:
loading a circulating cell-free nucleic acid derived from the blood of a fetus-bearing pregnant female into a sequencing device, or loading a modified variant of said nucleic acid into said sequencing device, said sequencing device produces a signal corresponding to the nucleotide bases of said nucleic acid;
generating sequence reads from said signal of said nucleic acid, optionally after transferring said signal to a system, by a system comprising one or more computing units, said one or more in said system The computing unit of includes memory and one or more processors,
A computing unit or combination of computing units in the system maps the reads of the sequence to a reference genome;
(a) using a microprocessor to generate a relationship between (i) local genomic bias estimates and (ii) bias frequencies for the sequence reads of the test sample, thereby generating a sample bias relationship; (said sequence reads are of circulating cell-free nucleic acid derived from said test sample;
the reads of said sequence are mapped against a reference genome),
(b) comparing said sample-biased relation to a reference-biased relation, thereby generating a comparison wherein said reference-biased relation is a combination of (i) a local genomic bias estimate for a reference and (ii) said bias frequency),
(c) normalizing the sequence read counts for the sample according to the comparison determined in (b), thereby reducing the sequence read bias for the sample; A method, including steps.
(Item 105)
105. The method of item 102, 103 or 104, wherein said normalizing in (c) comprises providing a normalized count number.
(Item 106)
106. The method of any one of items 102 to 105, wherein each of the local genomic bias estimates is determined by a process comprising the use of kernel density estimation.
(Item 107)
107. The method of any one of items 102-106, wherein each of the local genomic bias estimates for the reference bias relationship and the sample bias relationship is an indication of local bias content. .
(Item 108)
108. The method of item 107, wherein the local bias content is for polynucleotide segments of 5000 bp or less.
(Item 109)
109. The method of any one of items 102 to 108, wherein each of the local genomic bias estimates is determined by a process comprising using a sliding window analysis.
(Item 110)
110. The method of item 109, wherein the window is from about 5 consecutive nucleotides to about 5000 consecutive nucleotides and is slid from about 1 base to about 10 bases simultaneously in the sliding window analysis.
(Item 111)
110. The method of item 109, wherein said window is about 200 contiguous nucleotides and is slid about 1 base at a time in said sliding window analysis.
(Item 112)
(b) is a ratio comprising (i) a sample-biased related frequency and a reference-biased related frequency, each for each of said local genomic bias estimates, and (ii) a local genomic bias estimate; 112. A method according to any one of items 102 to 111, comprising generating a fitted relationship with
(Item 113)
113. The method of item 112, wherein the fitted relationship in (a) is obtained from weighted fitting.
(Item 114)
114. The method of any one of items 102-113, wherein each of said sequence reads for said sample is displayed in a binary format.
(Item 115)
115. The method of item 114, wherein said binary format for each read of said sequence comprises a chromosome to which said read maps and a chromosomal location to which said read maps.
(Item 116)
116. The method of item 115, wherein the binary format is in a 5-byte format comprising a 1-byte chromosomal ordinal number and a 4-byte chromosomal position.
(Item 117)
117. The method of any one of items 114 to 116, wherein said binary format is 50 times smaller than a Sequence Alignment/Map (SAM) format and/or approximately 13% smaller than a GZip format.
(Item 118)
118. Any one of items 102 through 117, wherein said normalizing in (c) comprises factorizing one or more features other than bias and normalizing read counts of said array The method described in .
(Item 119)
119. The method of item 118, wherein said factorization of one or more features is by a process that includes using a multivariate model.
(Item 120)
120. The method of item 119, wherein said processing including use of said multivariate model is performed by a multivariate module.
(Item 121)
121. The method of any one of items 118 to 120, wherein the sequence read counts are normalized according to the normalization in (c) and the factorization of the one or more features.
(Item 122)
(c) followed by generating a probability density estimate for each of said one or more portions comprising said count of reads of said sequence normalized in (c); or generating read densities for a plurality of parts or segments thereof.
(Item 123)
123. The method of item 122, wherein the probability density estimate is a kernel density estimate.
(Item 124)
124. A method according to item 122 or 123, comprising generating a read density profile for said genome or said segment thereof.
(Item 125)
125. The method of item 124, wherein said read density profile comprises said read density for said one or more portions of said genome or said segment thereof.
(Item 126)
126. The method of any one of items 122-125, comprising adjusting each of the lead densities for the one or more portions.
(Item 127)
127. The method of any one of items 122 to 126, wherein the one or more portions are filtered thereby providing filtered portions.
(Item 128)
128. The method of any one of items 122-127, wherein the one or more portions are weighted, thereby providing weighted portions.
(Item 129)
129. The method of item 128, wherein the one or more portions are weighted by eigenfunctions.
(Item 130)
130. The method of any one of items 102-129, wherein the local genomic bias estimate comprises a local GC density and the biased frequency comprises a GC biased frequency.
(Item 131)
(d) filtering the portion of the reference genome according to the read density distribution, thereby providing a read density profile for the test sample comprising the read density of the filtered portion;
wherein the read density comprises reads of sequences of circulating cell-free nucleic acids derived from test samples derived from pregnant females;
wherein the lead density distribution is determined for partial lead densities for a plurality of samples;
(e) adjusting said read density profile for said test sample according to one or more principal components obtained from a set of known euploid samples by principal component analysis, whereby a test sample comprising an adjusted read density; providing a profile;
(f) comparing the test sample profile to a reference profile, thereby providing a comparison;
(g) determining the presence or absence of a chromosomal aneuploidy for said test sample according to said comparison.
(Item 132)
132. The method of item 131, wherein the read density profile is adjusted in (b) by 1-10 principal components.
(Item 133)
132. The method of item 131, wherein the read density profile is adjusted in (b) by five principal components.
(Item 134)
wherein said one or more principal components adjust for one or more features in a read density profile, wherein said features are fetal sex, sequence bias, fetal fraction, bias correlated to DNase I sensitivity, entropy , repeat bias, chromatin structure bias, polymerase error rate bias, batch sequence bias, inverted repeat bias, PCR amplification bias, and hidden copy number variation of items 131 to 133. A method according to any one of paragraphs.
(Item 135)
135. The method of item 134, wherein the sequence bias comprises a guanine and cytosine (GC) bias.
(Item 136)
136. The method of any one of items 131-135, wherein said comparing comprises determining a level of significance.
(Item 137)
137. The method of item 136, wherein determining the level of significance comprises determining a p-value.
(Item 138)
138. The method of any one of items 131-137, wherein said reference profile comprises a read density profile obtained from a series of known euploid samples.
(Item 139)
139. The method of any one of items 131 to 138, wherein the reference profile comprises read densities of a filtered portion.
(Item 140)
140. The method of any one of items 131-139, wherein the reference profile comprises read densities adjusted according to the one or more principal components.
(Item 141)
141. Any one of items 136 to 140, wherein said level of significance indicates a statistically significant difference between said test sample profile and said reference profile and said presence of a chromosomal aneuploidy is determined described method.
(Item 142)
142. The method of any one of items 131-141, wherein said plurality of samples comprises a series of known euploid samples.
(Item 143)
143. The method of any one of items 131-142, wherein the lead density of a portion for the plurality of samples is a median lead density.
(Item 144)
144. The method of any one of items 131-143, wherein the read density of the filtered portion for the test sample is a median read density.
(Item 145)
145. The method of any one of items 138-144, wherein the read density profile for the reference profile comprises a median lead density.
(Item 146)
146. The method of any one of items 138-145, wherein the read densities for the test sample profile, the plurality of samples and the reference profile are determined according to a process comprising using kernel density estimation.
(Item 147)
147. The method of any one of items 144-146, wherein the test sample profile is determined according to the median read density for the test samples.
(Item 148)
148. The method of any one of items 145-147, wherein the reference profile is determined according to the median read density distribution for the references.
(Item 149)
149. The method of any one of items 131-148, comprising filtering portions of the reference genome according to an uncertainty measure for the read density distribution.
(Item 150)
149. The method of item 149, wherein the uncertainty measure is MAD.
(Item 151)
151. The method of any one of items 131-150, wherein said test sample profile is an indication of chromosomal dose for said test sample.
(Item 152)
152. The method of item 151, comprising comparing the chromosome dose for the test sample profile to the chromosome dose for the reference profile, thereby generating a chromosome dose comparison.
(Item 153)
153. The method of item 152, wherein determining the presence or absence of a chromosomal aneuploidy for the test sample is according to a comparison of the chromosome dose.
(Item 154)
Determining the presence or absence of a chromosomal aneuploidy for the test sample results in: 1 copy of the chromosome, 2 copies of the chromosome, 3 copies of the chromosome, 4 copies of the chromosome, 5 copies of the chromosome, 154. A method according to any one of items 131 to 153, comprising identifying the presence or absence of deletions of one or more segments of or insertions of one or more segments of a chromosome.
(Item 155)
A method for determining the presence or absence of aneuploidy for a sample, comprising:
(a) filtering portions of the chromosomes in the reference genome according to the read density distribution, thereby providing a read density profile for the test sample comprising the read densities of the filtered portions;
the read density comprises reads of sequences of circulating cell-free nucleic acids from test samples derived from pregnant females;
wherein the lead density distribution is determined for partial lead densities for a plurality of samples;
(b) adjusting said read density profile of chromosomes for said test sample according to one or more principal components obtained from a set of known euploid samples by principal component analysis, thereby comprising an adjusted read density; providing a test sample chromosomal profile;
(c) comparing said test sample chromosomal profile to a reference profile, thereby providing a comparison;
(d) determining the presence or absence of a chromosomal aneuploidy for said test sample according to said comparison.

Claims (22)

A method for determining the presence or absence of a chromosomal aneuploidy for a sample, comprising:
(a) filtering a portion of the reference genome according to the read density distribution, thereby providing a read density profile for the test sample comprising the read density of the filtered portion;
the read density comprises reads of sequences of circulating cell-free nucleic acids derived from a test sample derived from a subject;
wherein the lead density distribution is determined for partial lead densities for a plurality of samples;
(b) one or more principal components (i) obtained from a set of known euploid samples by principal component analysis and (ii) representing one or more biases in the read density profile; adjusting the read density profile for the test sample by subtracting from one or more read densities in the profile, thereby providing a test sample profile comprising the adjusted read density, wherein wherein multiple biases are excluded from the read density profile;
(c) comparing said test sample profile to a reference profile, thereby providing a comparison;
(d) determining the presence or absence of a chromosomal aneuploidy for said test sample according to said comparison;
A method, including The method of claim 1, wherein the read density profile is adjusted in (b) by 2-10 principal components. 2. The method of claim 1, wherein the read density profile is adjusted in (b) by five principal components. The one or more principal components represent multiple biases in the read density profile, wherein the biases are gender, sequence bias, bias correlated to DNase I sensitivity, entropy, repetitive sequence bias, chromatin structure bias. 4. The method of claim 1, 2 or 3, wherein the method is selected from bias, polymerase error rate bias, palindrome bias, inverted repeat bias, PCR amplification bias, and hidden copy number variation. 5. The method of claim 4, wherein the sequence bias comprises a guanine and cytosine (GC) bias. 6. The method of any one of claims 1-5, wherein said comparing comprises determining a level of significance. 7. The method of claim 6, wherein determining the level of significance comprises determining a p-value. The reference profile is
(i) read density profiles obtained from a series of known euploid samples;
(ii) the read density of the filtered portion; and/or
(iii) adjusted read density according to said one or more principal components;
8. A method according to any one of claims 1 to 7, comprising Claim 6 or Claim 7, or Claim 6, wherein said level of significance indicates a statistically significant difference between said test sample profile and said reference profile and said presence of a chromosomal aneuploidy is determined. 9. A method according to claim 8 when citing or 7. 10. The method of any one of claims 1-9, wherein the plurality of samples comprises a series of known euploid samples. (i) the read density of the portion for the plurality of samples is the median read density;
(ii) the read density of the filtered portion for the test sample is the median read density; and/or
(iii) the read density profile for the reference profile comprises a median read density;
11. A method according to any one of claims 1-10. 12. The method of any one of claims 8-11, wherein the read densities for the test sample profile, the plurality of samples, and the reference profile are determined according to a process comprising using kernel density estimation. Cite claim 11 or claim 11, wherein the test sample profile is determined according to the median read density for the test samples and the reference profile is determined according to the median read density distribution for the reference 13. The method of claim 12 in the case. 14. The method of any one of claims 1-13, comprising filtering portions of the reference genome according to an uncertainty measure for the read density distribution. 15. The method of claim 14, wherein the uncertainty measure is MAD. 16. The method of any one of claims 1-15, wherein the test sample profile is an indication of chromosomal dose for the test sample. 17. The method of claim 16, comprising comparing the chromosome dose for the test sample profile to the chromosome dose for the reference profile, thereby generating a chromosome dose comparison. 18. The method of claim 17, wherein determining the presence or absence of a chromosomal aneuploidy for the test sample is according to the comparison of chromosomal dose. Determining the presence or absence of a chromosomal aneuploidy for the test sample results in: 1 copy of the chromosome, 2 copies of the chromosome, 3 copies of the chromosome, 4 copies of the chromosome, 5 copies of the chromosome, 19. The method of any one of claims 1-18, comprising identifying the presence or absence of deletions of one or more segments of or insertions of one or more segments of a chromosome. 20. The method of any one of claims 1-19, comprising obtaining reads of the sequence prior to (a). A method for determining the presence or absence of a chromosomal aneuploidy for a sample, comprising:
(a) normalizing counts of sequence reads mapped to a reference genome, wherein said sequence reads are circulating cell-free nucleic acid reads from a test sample from a subject; , said normalization is
(1) generating a fitted relationship between (i) guanine and cytosine (GC) density and (ii) GC density frequency for reads of said sequence of said test sample, thereby generating a sample GC density relationship; wherein the sequence reads are mapped against the reference genome;
(2) comparing the sample GC density relationship with a reference GC density relationship, thereby generating a comparison;
said reference GC density relationship is between (i) a GC density and (ii) said GC density frequency for a reference;
(3) normalizing the sequence read counts for the sample according to the comparison determined in (2), thereby reducing the sequence read bias for the sample;
(b) filtering portions of the reference genome according to a read density distribution, thereby providing a read density profile for the test sample comprising the read densities of the filtered portions;
the read density comprises reads of sequences derived from the test sample;
wherein the lead density distribution is determined for partial lead densities for a plurality of samples;
(c) one or more principal components (i) obtained from a set of known euploid samples by principal component analysis and (ii) representing one or more biases in the read density profile; adjusting the read density profile for the test sample by subtracting from one or more read densities in the profile, thereby providing a test sample profile comprising the adjusted read density, wherein wherein multiple biases are excluded from the read density profile;
(d) comparing the test sample profile to a reference profile, thereby providing a comparison;
(e) determining the presence or absence of a chromosomal aneuploidy for said test sample according to said comparison;
A method, including A method for determining the presence or absence of a chromosomal aneuploidy for a sample, comprising:
(a) filtering portions of the chromosomes in the reference genome according to the read density distribution, thereby providing a read density profile for the test sample comprising the read densities of the filtered portions;
the read density comprises reads of sequences of circulating cell-free nucleic acids derived from a test sample derived from a subject;
wherein the lead density distribution is determined for partial lead densities for a plurality of samples;
(b) one or more principal components (i) obtained from a set of known euploid samples by principal component analysis and (ii) representing one or more biases in the read density profile; adjusting the read density profile of chromosomes for the test sample by subtracting from one or more read densities in the profile, thereby providing a test sample chromosome profile comprising the adjusted read densities; wherein a plurality of biases are excluded from said read density profile;
(c) comparing said test sample chromosomal profile to a reference profile, thereby providing a comparison;
(d) determining the presence or absence of a chromosomal aneuploidy for said test sample according to said comparison;
A method, including

Images